View
1
Download
0
Category
Preview:
Citation preview
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
ContentsMethodology....................................................................................................................................................2
Paradigm......................................................................................................................................................3
Scope............................................................................................................................................................4
Delimitations................................................................................................................................................4
A note on terminology..............................................................................................................................5
Self-selection............................................................................................................................................5
The experiments...........................................................................................................................................5
Framing the interventions........................................................................................................................6
The surveys...................................................................................................................................................6
Data set........................................................................................................................................................7
Expectations.................................................................................................................................................7
Data analysis.................................................................................................................................................8
Analysis.............................................................................................................................................................9
First grouping – all 4 interventions together................................................................................................9
Second grouping – each intervention separately.......................................................................................11
Make a list of five things you got done in your workday........................................................................11
Saying good morning..................................................................................................................................12
Write down 3 positive things from your workday......................................................................................15
Praise..........................................................................................................................................................16
Placebo.......................................................................................................................................................18
Recap..........................................................................................................................................................19
Implications....................................................................................................................................................20
Data............................................................................................................................................................20
Framing.......................................................................................................................................................20
Results........................................................................................................................................................21
Conclusion......................................................................................................................................................22
Recommendations..........................................................................................................................................22
Workshops..............................................................................................................................................22
References......................................................................................................................................................24
Page 1 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Methodology One of the main resources of an organization is its human capital i.e. its employees (Barney et al., 2001).
The motivation of the employees can be what determines the level of productivity and ultimately success
of the organization. In this respect, motivation can be understood as the process that initiates, guides and
maintains goal-directed behaviors. While early explanations of motivation have relied on instinct,
psychologists in recent years have suggested that individuals behave in certain ways depending on the
behavioral cues to which they are exposed (Hull, 1943; Tolman, 1948).
Motivation and being happy have been linked since the beginning of motivational theory. This study will
focus exclusively on the happy side of the link. The intent of this study is to make some headway into
thinking about how we talk about workplace happiness, and how measures to increase workplace
happiness has been implemented. Opinions on workplace happiness vary, some deem it creepy 1while
other have a much more positive outlook on bringing happiness into the workplace2. The act of
implementation has been the source of much controversy3, should companies be involved in the happiness
of their employees and if so – to what extend? However, the aim of this study is not to debate if workplace
happiness is worthwhile. The idea behind this study is to discern whether relative small interventions has
an effect on workplace happiness. This study aims to move away from the classical thinking in Danish
management that workplace happiness is something that is under prioritized and is driven, somewhat
reluctantly, by a yearly large intervention4. To move away from this thinking, this study proposes four small
interventions that can be done quickly and efficiently each day, with minimal interference in the “natural”
workday. Moreover, this study aims to explore the correlation between the proposed precise minimalistic
changes in behavior and workplace happiness. It seeks to examine the following hypothesis
How difficult is it to influence workplace happiness using only minor interventions?
The study also aims to answer the following questions:
What effect does compliance have on the effects of the interventions?
What effect does selfselection have on the effects of the interventions?
1 http://www.newrepublic.com/article/118804/happiness-officers-are-spreading-across-america-why-its-bad2 M. Seligman et al(2005); Positive Psychological Progress Emperical Validations of Interventions.3 http://www.newrepublic.com/article/118804/happiness-officers-are-spreading-across-america-why-its-bad4 Citation needed
Page 2 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
To what extent does the nature of the intervention have an effect on the result?
How big an influence did the time constraint have on the effects of the interventions?
The study will also work as a criticism of the idea that workplace happiness can be solved solely by
watching a short video, and doing short exercises. While these activities are not necessarily useless, their
effects are regularly overstated, often to increase sales or promote agendas. This study takes a hard look at
this practice to see what we can learn from it. Also, it happens to be a good source of self-critique.
The paper seeks to examine how far you can push workplace happiness using only small means.
ParadigmA major influence on any research design is the choice of a scientific paradigm. This section seeks to explain
the paradigm and the reasons for why it has been chosen. We take departure based on the following
definition of what a scientific paradigm is:
“A basic set of beliefs that guide action, whether of the everyday garden variety or action taken in
connection with a disciplined inquiry” (Guba, 1990).
This paper follows the objectivistic school of thought, whereby the ontological assumption on the nature of
reality is, that there exist one objective reality, which can be found through empirically testing questions
and hypothesis (Jonassen, 2011). This approach is typically based on a quantitative research method. There
exists a truth beyond the people who inhabit the world and it can be discovered.
The scientific paradigm used in this study is largely a positivistic view. A hypothesis has been proposed
which needs to be empirically and theoretically tested. In such, it implicates that there is a truth to be
found where our hypothesis is either true or false. By conducting a empirical study and reviewing the
prevailing theoretical literature, evidence for whether the hypothesis is true or false will be provided. This is
a positivistic view of the world where the truth is an objective entity, which will be discovered through
experimentation and research. This stands in contrast to the constructivist paradigm where there does not
exist one truth which can be found through empirical and theoretical testing (Jonassen, 2011). The truth is
observed between individuals, and is such a constructed truth. It is grounded in perception and
construction.
The thesis uses a quantitative research method where the focus is on one large empirical study in
conjunction with research on the prevailing literature on the different subjects related to the purpose of
this thesis. The empirical study use a quantitative research method in an effort to get a higher level of data
Page 3 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
validation. By examining the results of these studies and the prevailing theory it will be possible to answer
our research question.
ScopeThe study was limited to a run of approximately four weeks, and around 70 departments of various Danish
business participated. This means that approximately 1000 people used the interventions to promote
workplace happiness for four weeks. A series of videos was produced to aid in the articulation of the
intervention. All the external communication was done via the website: Arbejdsglædenu.dk. Five
interventions was made to conduct a classic A/B test, with one intervention designed as a placebo study. In
order to test whether or not the interventions would have an effect the workplace happiness of the
participants was measured three times during the study. The triple measurement of the participants
workplace happiness was deemed to be the maximum capability. It allowed for a sample of workplace
happiness in three separate stages; one before the interventions, one during the interventions and one
final after the conclusion of the study.
DelimitationsThere is a decent argument to be made for the questionnaires being inadequate. They were designed years
ago, and not with this kinds of experiment in mind. However, due to both time constraints and a desire for
this study to be somewhat backwards compatible, the questionnaires were left largely unchanged. In order
to get clearer results, a test study could have been made prior to the three main ones. This preliminary test
questionnaire could have helped in answering the following questions: did the participants use the
intervention, or a variant of it, before the study started. Was there a general commitment to the
participation of the study in the respective department.
These questions were later added to measurements 2 and 3, but they would have been much better served
in measurement 1. As the information gained could have been used to adjust the interventions, and
perhaps even lead to a better set of interventions. In the hypothesis above, we briefly touch upon
selfselection. Later on, we will go into a bit more depth with this, but an important point here is that we
could have designed 10-15 interventions and put them to a vote, in order to incorporate some form of
selfselection. We did not have time to even consider this, but it should have been discussed prior to the
study.
Because we are trying to uncover some of the mechanics associated with workplace happiness, complexity
is quite high, and it would have been a good idea to supplement these quantitative findings with qualitative
Page 4 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
interviews. Time and resources prevented us from doing this, but it could have added valuable knowledge
into some of the more process oriented issues that could have an effect on the study.
A note on terminologyThroughout this paper, the word workplace happiness is used. However, it is used to convey an
approximate translation of the Danish word Arbejdsglæde. The Danish word arbejdsglæde is not just about
being happy at work, it is also the happiness that comes from working(Kjerulf, 2010). Thus it is a subtle mix
of the satisfaction, desire and happiness gained from work and working. The study uses workplace
happiness as a direct translation of Arbejdsgæde.
Self-selectionSelf-selection is a broad term, in this study we primarily use it to describe and reflect on the fact that the participants of the study had no real input into the interventions. Similarly, it will be used to comment on how we gathered our data. We use self-selection to refer to the choice, or lack of choice within the study.
The experiments Designing the experiment was done using a framework articulated by Arbejdsglædnu / Woohoo Inc. The
experiment was quickly conceptualized into the five behavioral interventions mentioned above.
The design of the interventions came not only from a series of articles on positive psychology, but also
from Arbejdsglædenu’s product portfolio. The portfolio of Arbejdsglædenu have been build up the last 5
years, using a wide range of psychological theories(x reference) and a lot of trial and error. That is not to
say that the products developed by Arbejdsglædenu, and thus the stepping-stone for this experiement,
were unscientific. The products were developed using methods gained from scientific discoveries made in
positive psychology and current neuroscience. The interventions codifies some of the products proposed by
Arbejdsglædenu and seeks to validate them using a scientific method.
The interventions was made with the following intentions:
It needs to be easy
It needs to be quick
It needs to be something everyone can
do – it cannot involve a narrow skill
It has to make sense for the participant –
it has to involve an everyday activity
It needs be fairly easy to describe
It needs to be able to be performed daily
The argument behind all the intentions was to make sure the interventions could be performed accurately,
daily and with the highest chance of compliance.
Page 5 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
The five interventions chosen were:
1. Saying good morning5
2. Giving due praise and recognition6
3. Write down 3 positive things from your workday7
4. Write down 5 things you completed today – no matter how small8
5. Writing down the names of the last 3 people you spoke to ( Placebo)9
Each intervention was set up with its own webpage and video to guide and explain the purpose behind the
intervention and the study to the participants. Great care was taken to make sure each intervention was
presented in a homogenous fashion, not to mention that all information was displayed with clarity.
Framing the interventionsBecause of the need for the combination of this study to both have a scientific and a business value, the call
for participants was a unique affair. The participants were drawn mainly from old clients and other heavily
enfranchised individuals. In order to attract a sufficient number of participants the study was framed as a
series of “free stuff” from the participant point of view. This allowed us to gain
In order not to alienate the participants, the interventions were also presented in a way that made it seem
that the participants were already doing the interventions, and only needed to be more educated about it.
This was to seem too moralizing and, as with everything else on the website, the language was kept in a
positive frame.
The surveys The surveys were a series of multiple-choice questionnaires, with a few open ended questions. The were
build with the aim of quantifying workplace happiness. Rather than developing brand new questions for the
survey, the questions used regularly by ArbejdsglædeNu/Woohoo were modified to include a wider range
of responses. This was not done to be lazy, but to ensure that the new surveys were compatible with the
old ones. Also to ensure that this study was combatable with other projects being done simultaneously
there was an incentive to change as little as possible to the surveys.
5 http://arbejdsglaedenu.dk/hent-materialer/intervention-tghf/6 http://arbejdsglaedenu.dk/hent-materialer/intervention-ihds/7 http://arbejdsglaedenu.dk/hent-materialer/intervention-plnb/8 http://arbejdsglaedenu.dk/hent-materialer/intervention-uhgw/9 http://arbejdsglaedenu.dk/hent-materialer/intervention-gvfr/
Page 6 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Data setThe data set was acquired using a series of questionnaires powered by Google Drive. During the study, each
participant will have answered three questionnaires regarding their workplace happiness. The total pool of
participants in the first measurement was 1200 people, which provides amble statistical validity. There was
a significant drop in participants for the second and third measurements though. The second measurement
had 579 participants, where the third measurement had 560 participants. It seems that the drop of
participants was consistent past the first measurement. Luckily, the drop was consistent among the five
interventions, so that we were able to obtain data from all five of the interventions.
There were some minor differences in the questionnaires. Measurement two and three added some
additional questions, in order to test different parameters. Were we interested in uncovering additional
parameters such as: how compliant the test persons were and how much support there were to participate
in this test in their respective departments.
In order to ensure the validity of the data sets, they were subjected to both an ANOVA test and t-test with
two samples. This was to prove whether it would be possible to generalize the data obtained from this case
study.
Expectations In order to attempt to “debias” the conclusions presented within this study, the expectations we had
before attempting this study will be listed here.
All the four of the interventions was expected to have some form of positive effect. Also it was expected
that these interventions could produce some form of positive effect even on such a small timeframe. This
skewed the presentation of the data gained by this study, as we looked for these results. We expected the
placebo study to not significantly induce an increase within the different categories.
Page 7 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Data analysisBecause the questionnaires consisted of 10 questions all with a 1-7 scale. The focal characteristic of the
analysis is going to be the mean, that is the calculated average of the responses. In order to get a better
view of the data, we will organize it into groups. We will calculate and explore the change in mean, over the
3 measurements. These changes was grouped in two ways.
1. To get an overview of the general effects of the interventions we wanted to compare the general
mean of the 3 measurements to each other, this grouping was calculated using the following
parameters:
The first measurement was a simple mean calculation. For it to provide a baseline for a
comparison between the two other measurements, everything else was left as is
The second and third measurement was adjusted for compliance and the placebo effects
were removed. This meant that anyone who had answered lower than “2” on the
compliance question was filtered away, to get a clear idea of the actual effects of the
interventions.
2. To get a better understanding of which intervention was the most effective, and to see if there
would be a distinct difference of effect, the five interventions were separated. This is our second
grouping.
Like the first grouping, measurement 1 was kept as is, to form a baseline. Again this made it
easier to compare the three different measurements.
Because we were interested in the actual effects of the intervention, measurements 2 and
3 were adjusted for compliance. We filtered anyone who answered lower than “2” on the
compliance question away.
Each intervention was treated separately, all the participants who were given a specific
intervention was only pooled with people who had been given the same intervention.
These two groupings constitute largest body of evidence for our hypothesis. Both groupings as subjected
to an ANOVA and T test. As the aim of the study is to uncover whether or not it was possible to generalize
the findings obtained.
Analysis
Page 8 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
First grouping – all 4 interventions togetherPresented below are graphs detailing the development of the first grouping.
From this graph, we can make the following observations:
6/10 of the categories experiences a small increase in mean. Styker and Ros has the largest increase of all at 0,3.
Leder is the only one of the categories that experiences a small drop in mean, it drops from 5,6 to 5,5.
3/10 categories experience a small increase from measurement 1 to 2, but then a drop from measurement 2 to 3. This is an interesting phenomenon, and will be explored more later on.
Before we go into more depth on what these observations can mean, we should first look to the hypothesis presented in the introduction. As this graph gives a general overview on the effects of all the interventions, it is perfect for trying to prove our hypothesis. However, since we do not have the time or the resources to perform this experiment upwards of 100 times, we rely on statistical analysis to see if our results can be generalized.
To do this, we used ANOVA: Single Factor
Anova: Single Factor
Page 9 of 29Thomas Christensen & Alexander Kjerulf
Glad Energi Fysisk God Travlt Styrker Kolleger Leder Omsorg Ros3
3,5
4
4,5
5
5,5
6
6,5
5,3
4,9 5,0
5,5
4,7
5,0
5,9
5,6
5,2
4,5
5,3
5,0 5,0
5,5
4,8
5,2
6,0
5,65,5
4,8
5,4
5,0 5,1
5,5
4,7
5,3
5,9
5,55,3
4,8
Measurement 1,2 & 3
Måling nr. 1 Måling nr. 2 Målring nr. 3
Gen
nem
snit
værd
i for
sva
r
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
SUMMARYGroups Count Sum Average Variance
Measurement 1 10 51,64 5,164 0,193116Measurement 2 10 52,72 5,272 0,141507Measurement 3 10 52,46 5,246 0,116427
ANOVASource of Variation SS df MS F P-value F crit
Between Groups 0,063547 2 0,031773 0,21133 0,810833 3,354131Within Groups 4,05944 27 0,15035
Total 4,122987 29
The key thing to remember here, is that we are trying to disprove the null hypothesis of all 3 sets of data are equal. If we can disprove that, results from this study can be generalized.
To disprove the null hypothesis we need F > F crit. This has not been met, as 3,354131 F Crit > 0,21133 F. Therefore, we cannot disprove the null hypothesis. All 3 measurements are essentially “equal”, we cannot prove that the change in mean, does not come from variance. To be sure about this result, a T test: Two sample assuming equal variances was made.In this test measurement 1 and 3 were tested against each other. Measurement 2 only serves as a data tracking point for the interventions and thus omitted.
Measurement 1 Measurement 3Mean 5,164 5,246Variance 0,193116 0,116427Observations 10 10Pooled Variance 0,154771Hypothesized Mean Difference 0df 18t Stat -0,46607P(T<=t) one-tail 0,323377t Critical one-tail 1,734064P(T<=t) two-tail 0,646754t Critical two-tail 2,100922 Here we are interested in t Critical two-tail < t stat < t Critical two-tail. Obviously, this is not met here, which means that we cannot disprove our null hypothesis, both sets of data are equal. Thus, we cannot generalize based on this data. While there has been a small overall improvement in workplace happiness, within reasonable parameters, 95% chance, we cannot argue that this effect would persist if we were to replicate this study. However, before we dismiss the results of our study completely we will look at each of the interventions.
Page 10 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Second grouping – each intervention separatelyAs explained above, this section will examine the observations and statistical validity of each of the interventions. This means we will list the key observations made and then perform both ANOVA and T test. This section is likely going to get a bit repetitive, but do not worry afterwards we will get to the fun part of the analysis.
Make a list of five things you got done in your workday
From this graph, we can make the following observations:
7/10 categories experiences a small increase in mean. Leder is the only category that experiences a small drop in mean. 2/10 categories experience an increase in mean from measurement 1 to 2, but a drop from
measurement 2 to 3.
As we did above, we are going to perform both an ANOVA and T test, to see if we can generalize the data. Again we leave out measurement 2,
Anova: Single Factor
SUMMARYGroups Count Sum Average Variance
Page 11 of 29Thomas Christensen & Alexander Kjerulf
Glad Energi Fysisk God Travlt Styrker Kolleger Leder Omsorg Ros0
1
2
3
4
5
6
7
5,2
4,8 4,9
5,3
4,5
5,0
5,95,6
5,0
4,3
5,04,8
5,0
5,5
4,85,2
5,95,5 5,4
4,5
5,45,1
4,9
5,7
4,7
5,4
5,9
5,4 5,4
4,6
Måling nr. 1 Måling nr. 2 Målring nr. 3
Gen
nem
snit
værd
i for
sva
r
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Measurement 1 10 50,5 5,05 0,229444Measurement 3 10 52,8 5,28 0,197333
ANOVASource of Variation SS df MS F P-value F crit
Between Groups 0,2645 1 0,2645 1,239521 0,280211 4,413873Within Groups 3,841 18 0,213389
Total 4,1055 19
To disprove the null hypothesis we need F > F crit. This has not been met, as 4,413873F Crit > 1,239521F. Therefore, we cannot disprove the null hypothesis. Both measurements are essentially “equal”; we cannot prove that the change in mean does not come from variance.
t-Test: Two-Sample Assuming Equal Variances
Measurement 1 Measurement 3Mean 5,05 5,28Variance 0,229444 0,197333Observations 10 10Pooled Variance 0,213389Hypothesized Mean Difference 0df 18t Stat -1,11334P(T<=t) one-tail 0,140105t Critical one-tail 1,734064P(T<=t) two-tail 0,280211t Critical two-tail 2,100922
Here we are interested in t Critical two-tail < t stat < t Critical two-tail. Obviously, this is not met here, which means that we cannot disprove our null hypothesis, both sets of data are equal. Thus, we cannot generalize based on this data.
Page 12 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Saying good morning
Observations:
3/10 categories experience a small increase in mean. 6/10 categories experience a small decrease in mean. Energi and Omsorg experience a small increase in mean, from measurement 1 to 2, but a drop in
measurement 3, this decrease drops the two categories below their original starting point.
Anova: Single Factor
SUMMARY
Groups Count Sum Average Variance
Measurement 1 10 51,1 5,11 0,205444
Measurement 3 10 50,6 5,06 0,116
ANOVA
Source of Variation SS df MS F P-value F crit
Page 13 of 29Thomas Christensen & Alexander Kjerulf
Glad Energi Fysisk God Travlt Styrker Kolleger Leder Omsorg Ros0
1
2
3
4
5
6
7
5,2
4,9 4,8
5,4
4,9 4,9
5,95,6
5,2
4,3
5,25,0
4,7
5,5
4,95,2
5,95,5 5,4
4,6
5,1
4,74,9
5,3
4,7
5,2
5,75,3
5,1
4,6
Måling nr. 1 Måling nr. 2 Målring nr. 3
Gen
nem
snit
værd
i for
sva
r
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Between Groups 0,0125 1 0,0125 0,077774 0,783516 4,413873
Within Groups 2,893 18 0,160722
Total 2,9055 19
To disprove the null hypothesis we need F > F crit. This has not been met, as 4,413873 F Crit 0,077774 F. Therefore, we cannot disprove the null hypothesis. Both measurements are essentially “equal”; we cannot prove that the change in mean does not come from variance.
t-Test: Two-Sample Assuming Equal Variances
Measurement 1 Measurement 3Mean 5,11 5,06Variance 0,205444 0,116Observations 10 10Pooled Variance 0,160722Hypothesized Mean Difference 0df 18t Stat 0,27888P(T<=t) one-tail 0,391758t Critical one-tail 1,734064P(T<=t) two-tail 0,783516t Critical two-tail 2,100922
Here we are interested in t Critical two-tail < t stat < t Critical two-tail. Obviously, this is not met here, which means that we cannot disprove our null hypothesis, both sets of data are equal. Thus, we cannot generalize based on this data.
Page 14 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Write down 3 positive things from your workday
In this graph, we can make the following observations:
8/10 categories experience a small gain in mean. 2/10 categories experience a small drop in mean.
Anova: Single Factor
SUMMARYGroups Count Sum Average Variance
Measurement 1 10 52,2 5,22 0,164Measurement 3 10 54,1 5,41 0,141
ANOVASource of Variation SS df MS F P-value F crit
Between Groups 0,1805 1 0,1805 1,183607 0,290976 4,413873Within Groups 2,745 18 0,1525
Total 2,9255 19
Again,to disprove the null hypothesis we need F > F crit. This has not been met, as 4,413873 F Crit > 1,183607F. Therefore, we cannot disprove the null hypothesis
Page 15 of 29Thomas Christensen & Alexander Kjerulf
Glad Energi Fysisk God Travlt Styrker Kolleger Leder Omsorg Ros0
1
2
3
4
5
6
7
5,4
4,9 4,9
5,5
4,85,1
5,95,7
5,3
4,7
5,45,2 5,1
5,7
4,8
5,3
6,1
5,5 5,6
4,9
5,5 5,4 5,45,8
4,7
5,2
6,0
5,5 5,6
5,0
Måling nr. 1 Måling nr. 2 Målring nr. 3
Gen
nem
snit
værd
i for
sva
r
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
t-Test: Two-Sample Assuming Equal Variances
Måling 1 Måling 3Mean 5,22 5,41Variance 0,164 0,141Observations 10 10Pooled Variance 0,1525Hypothesized Mean Difference 0df 18t Stat -1,087936835P(T<=t) one-tail 0,145488157t Critical one-tail 1,734063607P(T<=t) two-tail 0,290976314t Critical two-tail 2,10092204
Here we are interested in t Critical two-tail < t stat < t Critical two-tail. We fail to disprove our null hypothesis.
Praise
Page 16 of 29Thomas Christensen & Alexander Kjerulf
Glad Energi Fysisk God Travlt Styrker Kolleger Leder Omsorg Ros0
1
2
3
4
5
6
7
5,4
5,05,3
5,6
4,7
5,2
6,1
5,75,3
4,6
5,5
5,1 5,2
5,6
4,8
5,2
6,15,9
5,55,2
5,6
5,2 5,2
5,7
4,8
5,4
6,0
5,6 5,55,1
Måling nr. 1 Måling nr. 2 Målring nr. 3
Gen
nem
snit
værd
i for
sva
r
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Here we can make the following observations:
7/10 categories experience a small increase in mean. 3/10 categories experience a small decrease in mean. The Leder categorie experiences a small boost from measurement 1 to 2, but then a drop from
measurement 2 to 3.
Anova: Single Factor
SUMMARYGroups Count Sum Average Variance
Måling 1 10 52,9 5,29 0,205444Måling 3 10 54,1 5,41 0,118778
ANOVASource of Variation SS df MS F P-value F crit
Between Groups 0,072 1 0,072 0,44414 0,513583 4,413873Within Groups 2,918 18 0,162111
Total 2,99 19
Again, to disprove the null hypothesis we need F > F crit. This has not been met, as 4,413873 F Crit > 0,44414 F. Therefore, we cannot disprove the null hypothesis
t-Test: Two-Sample Assuming Equal Variances
Måling 1 Måling 3Mean 5,29 5,41Variance 0,205444 0,118778Observations 10 10Pooled Variance 0,162111Hypothesized Mean Difference 0df 18t Stat -0,66644P(T<=t) one-tail 0,256792t Critical one-tail 1,734064P(T<=t) two-tail 0,513583t Critical two-tail 2,100922
Here we are interested in t Critical two-tail < t stat < t Critical two-tail. We fail to disprove our null hypothesis.
Page 17 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
PlaceboFor reference, we have included the placebo study.
As as
Glad Energi Fysisk God Travlt Styrker Kolleger Leder Omsorg Ros4
4,5
5
5,5
6
6,5
5,6
5,1 5,1
5,6
5,15,2
6,1
5,7
5,5
4,5
5,6
5,05,1
5,7
4,9
5,2
6,2
5,7
5,6
5,0
5,7
5,0
5,3
5,6
4,7
5,3
6,1
5,9
5,6
4,8
Måling nr. 1 Måling nr. 2 Målring nr. 3
Gen
nem
snit
værd
i for
sva
r
As with the previous interventions, we are going to run an ANOVA and t test. This time, however, we are hoping to confirm our null hypothesis.
Anova: Single Factor
SUMMARYGroups Count Sum Average Variance
Måling 1 10 53,5 5,35 0,196111Måling 3 10 54 5,4 0,215556
ANOVASource of Variation SS df MS F P-value F crit
Between Groups 0,0125 1 0,0125 0,060729 0,808136 4,413873Within Groups 3,705 18 0,205833
Total 3,7175 19
As F > F crit is not mere here, we can confirm our null hypothesis.
Page 18 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
t-Test: Two-Sample Assuming Equal Variances
Måling 1 Måling 3Mean 5,35 5,4Variance 0,196111 0,215556Observations 10 10Pooled Variance 0,205833Hypothesized Mean Difference 0df 18t Stat -0,24643P(T<=t) one-tail 0,404068t Critical one-tail 1,734064P(T<=t) two-tail 0,808136t Critical two-tail 2,100922
Like before, we are interested in t Critical two-tail < t stat < t Critical two-tail. This is not obtained, so we cannot disprove our null hypothesis.
RecapWhile we are by no means done with our analysis, we have just gone through some rather boring mathematical proofs. Here is a recap.
Neither of the two groupings gave us data that we could use to extrapolate what would happen if we were to repeat the study. At every single parameter, we were unable to disprove our null hypothesis.
The First grouping showed us a marginal overall increase in workplace happiness. In the Second grouping, three of the four intervention showed a marginal increase in workplace
happiness. The “good morning” intervention in the Second grouping showed very little movement across the
three measurements. The category of “Leder” behaved very differently than the other categories. “Leder” would
experience a drop in mean, consistently across both the First and Second grouping. The category of “ros” is extremely volatile, even in the placebo study it shows a great increase.
Therefore it is quite likely that participants make some heuristic leap in answering the “ros” question, as there is no evidence to support that all of the interventions would that effect on praise.
Page 19 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Implications
DataThe two obvious implications of the study are the fact the while we were unable to obtain data that could be used to generalize, the study did have a positive effect on the participants. These seem to contradict one another. How can the study have an effect, however marginal, and then end up with results that cannot be used to generalize? As stated in the introduction above, since we do not have the resources to perform the study 100 or so times, we use statistics to simulate it instead. This means that we have to acknowledge two conclusions from this study: 1. the study did have an effect and 2. we cannot conclude that we can replicate this effect in future.
Three of the four interventions showed an increase in the majority of the 10 parameters, ranging between 7/10 and 8/10, however the intervention of “saying Good Morning” only showed an increase in 3/10 parameters. This is likely because people participating in the intervention was already saying good morning to each other. Which means that we did not really observe any changes. When we asked the participants if they used the intervention prior to the start, the majority of the participants indicated that they already were saying good morning to each other every day. Making this particular intervention rather uninteresting.
As mentioned above, the category of “leder” experienced a universal drop across all of the intervention. When we combine this with the sharp drop in participants from measurement 2, along with the inconclusive results from above, it is not unreasonable to argue that some top down tensions has influenced the results. This will be explored more fully below.
FramingBefore we proceed with the implication of the results, we have to discuss some unusual aspect of the framing of the study. We set out to frame our study to be both beneficial for the people involved, but also to be valuable to sign up for. We wanted to make people who did not sign up feel they were missing out on something. As mentioned in the introduction, coupled with the ability to sign up your entire department, this ensured we got A LOT OF participants. In the end, we ended up in a very interesting place, as we had somewhat unwillingly created something that closely resembled the worst characteristics of a corporate action towards workplace happiness. With this in mind, we set out to make the best of it, the blatant disregard of self-selection made it so we could actually use our data to test what happens when you do workplace happiness in top down management.
There is an argument for our sample size to be skewed towards people who have a vested interest in workplace happiness, and thus our study would be more successful compared to a regular top down management decision. However, the layman’s concept of workplace happiness is that much of a platitude that when asked, everybody would have a vested interest in this.
Therefore, we have ended up with a double framing of our study. The first one is a classic “loss frame” as discussed above. The second one is the top down corporate “do these exercises and be happy” frame. The first one is targeted at people viewing the study, and are not yet participants. Adversely, the second framing is an internal framing, and is likely only encountered by us.
Page 20 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
ResultsThe main question that we should ask is “what conclusion we can draw from a study with largely inconclusive statistical data?”
We have previously discussed the expectations we had prior to starting the study, in the following section we are going to examine how the observations made from the interventions compare.
Because of its broad nature, there was an expectation that “glad” would experience a small increase in all of the interventions. Also, we expected there to be little to no adjustment in the “fysisk” category, as none of the interventions should have had any bearing on the participant, it did however. The category of “leder” was also not expected to experience a significant change, as none of the interventions directly involved the employer/employee relation. However, as stated in the recap the category of “leder” actually dropped in all the interventions.
It is important to take a moment to reflect on this. The way the signup worked was that one person, usually the leader of a department, could sign up the entire department. Because we included 3 workplace happiness measurements, there is an argument that this made participation of the study disproportionately more attractive to employers/leader of the department. What we failed to consider, was that a number of participants signed up because their respective leaders wanted workplace happiness measurements. This could account for the steady drop in the category “leder”. Some of the participants might be reacting negatively to the fact that their leader/employer signed them up for something they did not want to participate in. This leads into the idea of self-selection, a concept we mention it briefly in the introduction.
None of the participants had any say in what intervention they wanted to perform, which meant that if they felt they experienced a workplace happiness related problem they did not have an option to choose an intervention they felt could help them. The results is marred by a double disengagement of self-selection. As the participants: 1. had no choice in what intervention they were going to perform and 2. they had no choice in the sign up. No wonder the rate of questionnaire answers almost dropped in half from measurement 1 to 2.
As mentioned in the introduction, the rate of answers stayed consistence through measurement 2 and 3. Which could argue that the effect of self-selection lessened over the study, and once people got going they accepted the intervention and generally kept their efforts up.
Page 21 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
ConclusionTo conclude, we did three measurements over the course of 4 weeks to map fluctuations in our questionnaire. Our results strongly indicated that there were no significant change in workplace happiness during the 4 weeks of testing. A combination of a lack of self-selection and encouragement from peers and leaders lead to this result. Thus, we were unable to disprove the mathematical null hypothesis put forth in the analysis section.
To reiterate, it is extremely difficult to influence workplace happiness using only these small methods. There are numerous factors that get in the way, so that while the scientific argument behind the interventions is valid, there is simply too far from these generalizable methods to where the problem is. Doing top down “happiness exercises” does not have a meaningful impact on workplace happiness, even if on average each participant in the study experiences an increase across the questionnaire.
Our data suggest that self-selection has a large influence on the success of implementing interventions to help with workplace happiness. This is linked to the exact nature of the interventions, e.g. it was almost impossible to test “saying good morning” as the majority already performed something similar. Also, getting either people to write down what they got done today or positive things that happened to them during the day, can be very disruptive.
We can make up some general interventions that on the surface have an effect on workplace happiness, in that they constitute positive habits. Because they are positive habits, they are better suited for maintaining good workplace happiness, but because they are so general they are unsuited for change.
Recommendations With the above in mind, what would be the best way to implement initiatives or interventions to increase workplace happiness. The first step is to realize that workplace happiness is more complex than you might expect. This complexity comes not just from the fact that happiness in and of itself is a complex neurological phenomenon, but also from the nature of workplace happiness. As we discuss earlier, workplace happiness is really a combination of many factors. That is why the 4 broad interventions were chosen. Each were meant to affect a different aspect of workplace happiness, likewise the questionnaire has 10 categories to better track movement in workplace happiness.
Rather than approach workplace happiness with a top down idea, we should approach it from the bottom up. This allows us to work with self-selection, rather than against it. Approaching each workplace happiness related encounter as a case study would allow us to break up the workplace happiness issue into smaller chunks and design interventions that suit them. Not to mention design interventions that work with the participants workday, rather than just have interventions that exist on top of the workday.
WorkshopsWorkplace happiness is inherently complex. There is no easy solutions, especially if you want to use self-selection as a tool. Rather than trying to force good habits on participants using a top down method, it
Page 22 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
seems more prudent to work bottom up, with determined problems. Otherwise, you very easily end up playing the role of the “happiness dictator”, where you superimpose these otherwise beneficial habits on people. Workshops rather than lectures seem to work much better. Naturally, there is an issue of time. Further studies could determine whether some form of “micro workshops” could be feasible. The idea is that habits, even beneficial ones, do not work in a vacuum.
ReferencesNot all of the references listed below are quotes directly above. Instead, they have in some form or another provided a theoretical foundation for the study, which is why they are listed here. Many of the articles helped us shape the interventions presented above.
Page 23 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
About Neuroscience. (n.d.). Society for Neuroscience. Retrieved February 23, 2014, from
http://www.sfn.org/about/about-neuroscience
Baars, B. J., & Gage, N. M. (2013). Fundamentals of cognitive neuroscience: a beginner's guide.
Amsterdam: Academic Press.
Bargh, J. A., & Chartrand, T. L. (2000). Studying the Mind in the Middle: A Practical Guide to
Priming and Automaticity Research. Handbook of Research Methods in Social Psychology.
Barney, J. B., Wright, M., & Ketchen, D. J. (2001). The resource-based view of the firm: ten years
after 1991 . New York, NY: Pergamon.
Bogue, R. L. (2005, April 25). Use S.M.A.R.T. goals to launch management by objectives plan.
TechRepublic. Retrieved February 23, 2014, from http://www.techrepublic.com/article/use-
smart-goals-to-launch-management-by-objectives-plan/
Bryman, A., & Bell, E. (2011). Business research methods (3rd ed.). Cambridge. Oxford University
Press.
Case Study #1: Innovation - The Candle Problem. (n.d.). Asset. Retrieved March 1, 2014, from
http://2012e.igem.org/Team:Arizona_State_E/Candle
Chartrand, T. L., & Bargh, J. A. (1996). Automatic activation of impression formation and
Page 24 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
memorization goals: Nonconscious goal priming reproduces effects of explicit task
instructions. Journal of Personality and Social Psychology.
Chartrand, T. L., van Baaren, R. B., & Bargh, J. A. (2006). Linking Automatic Evaluation to Mood
and Information Processing Style: Consequences for Experienced Affect, Impression
Formation, and Stereotyping. Journal of Experimental Psychology.
Classical Conditioning Overview. (n.d.). Retrieved February 23, 2014, from
http://dakota.fmpdata.net/PsychAI/PrintFiles/ClassCondOvrvw.pdf
Coffey, J. O. (2009, July 1). Why Happy Employees Determine Your Success. Executive Travel.
Retrieved February 21, 2014, from http://www.executivetravelmagazine.com/articles/why-
happy-employees-determine-your-success
Connolly, J., & Viswesvaran, C. (2000). The role of activity in job satisfaction: a meta-
analysis. Psychology Department, Florida International University, Miami.
Diener, E., & Biswar-Diener, R. (2002). Will Money Increase Subjective Well-being? Published by
Springer.
Dooley, R. (2006). Priming the Customer. Neuromarketing RSS. Retrieved March 1, 2014, from
http://www.neurosciencemarketing.com/blog/articles/priming-the-customer.htm
Dooley, R. (2007, March 1). Priming by Order. Neuromarketing RSS. Retrieved February 23, 2014,
from http://www.neurosciencemarketing.com/blog/articles/priming-by-order.htm
Drahota, A., Costall, A., & Reddy, V. (2008). The vocal communication of different kinds of
smile. Speech Communication.
Eriksson, P., & Kovalainen, A. (2008). Qualitative methods in business research. Los Angeles.
Sage.
Guay, F., Chanal, J., Ratelle, C. F., Marsh, H. W., Larose, S., & Boivin, M. (2010). Intrinsic,
identified, and controlled types of motivation for school subjects in young elementary
Page 25 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
school children. British Journal of Educational Psychology.
Glimcher, P., Camerer, C., Fehr, E., & Poldrack, R. (2009). The Study of Emotion in
Neuroeconomics: Decision Making and the Brain (p. 233). London. Elsevier.
Guba, E. G. (1990). The Paradigm dialog. Newbury Park, California. Sage Publications.
Hamann, S. B., & Squire, L. R. (1997). Intact perceptual memory in the absence of conscious
memory. Behavioral Neuroscience.
Hanks, A. S., Just, D. R., Smith, L. E., & Wansink, B. (2012). Healthy convenience: nudging
students toward healthier choices in the lunchroom. Journal of Public Health.
Higgins, E., & Chaires, W. M. (1980). Accessibility of interrelational constructs: Implications for
stimulus encoding and creativity. Journal of Experimental Social Psychology.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category Accessibility and Impression
Formation. Journal of Experimental Social Psychology.
Hull, C. L. (1943). Principles of behavior, an introduction to behavior theory. New York, London:
D. Appleton-Century Company, incorporated.
Hultén, B. (2011). Sensory marketing: the multi-sensory brand-experience concept. European
Business Review,23(3)..
Jespersen, A., & Hansen, P. (n.d.). Nudge theory: The mechanics of the brain. iNudgeYou RSS.
Retrieved March 13, 2014, from http://www.inudgeyou.com/nudge-theory-the-mechanics-
of-the-brain/
Jonassen, D. H. (1991). Objectivism versus constructivism: Do we need a new philosophical
paradigm? Educational Technology Research and Development.
Kahneman, D. (2003). A Perspective on Judgment and Choice: Mapping Bounded Rationality.
American Psychologist.
Page 26 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Koivumaa-Honkanen, H., Koskenvuo, M., Honkanen, R. J., Viinamki, H., Heikkil, K., & Kaprio, J.
(1999). Life dissatisfaction and subsequent work disability in an 11-year follow-
up. Psychological Medicine,34(2), 221-228.
Lamar, D. D., & Laney, B. (n.d.). How to keep your employees happy and productive. How to keep
your employees happy and productive. Retrieved February 23, 2014, from
http://www.reliableplant.com/Read/18833/how-to-keep-your-employees-happy-productive
Lashley, K. S., & Jeffress, L. A. (1951). The problem of serial order in behavior. Pasadena.
California Institute of Technology.
Malone, J. C. (1990). Theories of Learning - A historical approach. Belmont, California.
Wadsworth Publishing Company.
Marteau, T. (2011). Judging nudging: can nudging improve population health? BMJ, 342, 263-265.
Milner, B., Squire, L. R., & Kandel, E. R. (1998). Cognitive Neuroscience and the Study of
Memory. Neuron, 20(3), 445-468.
Owen, A. M., Hampshire, A., Grahn, J. A., Stenton, R., Dajani, S., Burns, A. S., et al. (2010).
Putting Brain Training to the Test. Nature.
Poldrack, R. (2006). Can Cognitive Processes Be Inferred From Neuroimaging Data? Trends in
Cognitive Sciences, 10(2), 59-63.
Rokade, D. P. (2011). Release of Endomorphin Hormone and Its Effects on Our Body and Moods:
A Review. International Conference on Chemical, Biological and Environment Sciences.
Page 27 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
Rustichini, A. (2009). Neuroeconomics: what have we found, and what should we search for.
Current Opinion in Neurobiology, 19(6), 672-677.
Santos, M. (n.d.). Simple Maths in a Complex Brain: the probability of two equal states of mind.
Retrieved February 23, 2014, from http://jcienciascognitivas.home.sapo.pt/06-
01_vieira.html
Schacter, D. L., & Tulving, E. (1994).Memory systems 1994. Cambridge, Mass.: MIT Press.
Segal, S. J., & Cofer, C. N. (1960). The effect of recency and recall on word association. American
Psychologist, page 451.
Shantz, A., & Latham, G. P. (2009). An exploratory field experiment of the effect of subconscious
and conscious goals on employee performance. Elsevier.
Sherry, D. F., & Schacter, D. L. (1987). The evolution of multiple memory systems. Psychological
Review, page 440.
Springs, M. (2013, November 14). Why Happy Employees Are Your Key to Successful Branding.
Entrepreneur. Retrieved February 23, 2014, from
http://www.entrepreneur.com/article/229924
Squire, L. (2004). Memory Systems of the Brain: A Brief History and Current
Perspective. Neurobiology of Learning and Memory, 82(3), 171-177.
Srull, T. K., & Wyer, R. S. (1979). The Role of Category Accessibility in the Interpretation of
Information about Persons: Some Determinants and Implications. Journal of Personality
and Social Psychology.
Thaler, R. H., & Sunstein, C. R. (2008). Nudge: improving decisions about health, wealth, and
happiness. New Haven, Conn.: Yale University Press.
Ven, A. H. (1999). The innovation journey. New York: Oxford University Press.
Page 28 of 29Thomas Christensen & Alexander Kjerulf
Thomas Christensen & Alexander Kjerulf Woohoo Inc. January 2015
West, R. F., & Toplak, M. E. (2011). Individual differences as essential components of heuristics
and biases research. Psychology Press.
Wiggs, C. L., & Martin, A. (1998). Properties and mechanisms of perceptual priming. Current
Opinion in Neurobiology.
Williams, R. (1988). The Control of Neuron Number. Annual Review of Neuroscience, 11(1), 423-
453.
Wright, T. A., & Staw, B. M. (1999). Affect and favorable work outcomes: two longitudinal tests of
the happy-productive worker thesis . Journal of Organizational Behavior , 20 (1), 1-23.
Page 29 of 29Thomas Christensen & Alexander Kjerulf
Recommended