8
mental retardation. Other scales depend upon self- report (Ramirez & Kratochwill, 1997). Depression is one disorder frequently assessed by such scales. The fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (American Psy- chiatric Association [APA], 1994) reports that the rate of depression in the general population is “4.5% to 9.3% for females and 2.3% to 3.2% for males” (p. 229). Singh et al. (1991), in a review of the research, sug- gest the prevalence rate of depression in persons with mental retardation in a clinical sample is 8%. Depression (and psychiatric disorders in general) in persons with mental retardation is probably due to multiple interacting factors, including a lack of social support, poor social skills, the lack of a stimulating en- vironment, brain damage, medical and physical prob- lems, and the inability of these persons to think adaptively about themselves and their worlds. These complex and interacting factors make it difficult to tease out unique contributing factors. Pawlarcyzk and Beckwith (1987) reviewed reports of depression in persons with mental retardation in order to assess the applicability of DSM-III depression INTRODUCTION Singh, Sood, Sonenklar, and Ellis (1991) report that at least 50% of institutionalized persons with men- tal retardation have “at least one identifiable psychi- atric disorder,” and of those children with mental retardation who are living in the community, 20–35% will have a “diagnosable mental illness” (p. 422). To address the needs of these individuals, professionals are developing assessment and treatment methods spe- cific to persons with mental retardation. For example, the Diagnostic Assessment for the Severely Handi- capped (Matson, Gardner, Coe, & Sovner, 1991), the Psychopathology Inventory for Mentally Retarded Adults (Matson, Kazdin, & Senatore, 1984), and the Aberrant Behavior Checklist (Aman & Singh, 1986; Bihm & Poindexter, 1991) are completed by an infor- mant and assess a variety of disorders in persons with Depression, Anxiety, and Relevant Cognitions in Persons with Mental Retardation Elizabeth Glenn, Elson M. Bihm, 1,2 and William J. Lammers 1 We assessed depression, anxiety, and relevant cognitions in persons with mental retardation by administering modified versions of the Reynolds Child Depression Scale, the Beck Anxi- ety Inventory, the Automatic Thoughts Questionnaire, and the Cognitions Checklist to 46 per- sons with borderline to moderate mental retardation. Consistent with research with other groups, self-reports of depression and anxiety were highly correlated (r 5 .74) in these individuals, and cognitions were strong predictors of negative affect. Subscales measuring cognitions re- lated to depression and anxiety were also highly related, limiting the “cognitive-specificity” hypothesis. Hierarchical multiple regression analyses offered mixed support for cognitive- specificity. We discuss the implications of these findings for the cognitive and affective as- sessment of persons with intellectual limitations. KEY WORDS: Mental retardation; anxiety; depression; comorbidity; dual diagnosis; cognitions. Journal of Autism and Developmental Disorders, Vol. 33, No. 1, February 2003 (©2003) 69 0162-3257/03/0200-0069/0 © 2003 Plenum Publishing Corporation 1 University of Central Arkansas, Conway, Arkansas. 2 Department of Psychology and Counseling, University of Central Arkansas, Conway, AR 72035; Tel.: 501-450-5417; e-mail: [email protected]

Depression, Anxiety, and Relevant Cognitions in Persons with Mental Retardation

Embed Size (px)

Citation preview

mental retardation. Other scales depend upon self-report (Ramirez & Kratochwill, 1997).

Depression is one disorder frequently assessed bysuch scales. The fourth edition of the Diagnostic andStatistical Manual of Mental Disorders(American Psy-chiatric Association [APA], 1994) reports that the rateof depression in the general population is “4.5% to9.3% for females and 2.3% to 3.2% for males” (p. 229).Singh et al. (1991), in a review of the research, sug-gest the prevalence rate of depression in persons withmental retardation in a clinical sample is 8%.

Depression (and psychiatric disorders in general)in persons with mental retardation is probably due tomultiple interacting factors, including a lack of socialsupport, poor social skills, the lack of a stimulating en-vironment, brain damage, medical and physical prob-lems, and the inability of these persons to thinkadaptively about themselves and their worlds. Thesecomplex and interacting factors make it difficult totease out unique contributing factors.

Pawlarcyzk and Beckwith (1987) reviewed reportsof depression in persons with mental retardation inorder to assess the applicability of DSM-III depression

INTRODUCTION

Singh, Sood, Sonenklar, and Ellis (1991) reportthat at least 50% of institutionalized persons with men-tal retardation have “at least one identifiable psychi-atric disorder,” and of those children with mentalretardation who are living in the community, 20–35%will have a “diagnosable mental illness” (p. 422). Toaddress the needs of these individuals, professionalsare developing assessment and treatment methods spe-cific to persons with mental retardation. For example,the Diagnostic Assessment for the Severely Handi-capped (Matson, Gardner, Coe, & Sovner, 1991), thePsychopathology Inventory for Mentally RetardedAdults (Matson, Kazdin, & Senatore, 1984), and theAberrant Behavior Checklist (Aman & Singh, 1986;Bihm & Poindexter, 1991) are completed by an infor-mant and assess a variety of disorders in persons with

Depression, Anxiety, and Relevant Cognitions in Personswith Mental Retardation

Elizabeth Glenn, Elson M. Bihm,1,2 and William J. Lammers1

We assessed depression, anxiety, and relevant cognitions in persons with mental retardationby administering modified versions of the Reynolds Child Depression Scale, the Beck Anxi-ety Inventory, the Automatic Thoughts Questionnaire, and the Cognitions Checklist to 46 per-sons with borderline to moderate mental retardation. Consistent with research with other groups,self-reports of depression and anxiety were highly correlated (r 5 .74) in these individuals,and cognitions were strong predictors of negative affect. Subscales measuring cognitions re-lated to depression and anxiety were also highly related, limiting the “cognitive-specificity”hypothesis. Hierarchical multiple regression analyses offered mixed support for cognitive-specificity. We discuss the implications of these findings for the cognitive and affective as-sessment of persons with intellectual limitations.

KEY WORDS: Mental retardation; anxiety; depression; comorbidity; dual diagnosis; cognitions.

Journal of Autism and Developmental Disorders, Vol. 33, No. 1, February 2003 (©2003)

690162-3257/03/0200-0069/0 © 2003 Plenum Publishing Corporation

1 University of Central Arkansas, Conway, Arkansas.2 Department of Psychology and Counseling, University of Central

Arkansas, Conway, AR 72035; Tel.: 501-450-5417; e-mail:[email protected]

70 Glenn, Bihm, and Lammers

criteria to this population. They found that for indi-viduals with mild and moderate mental retardation,these criteria may be useful and appropriate for iden-tifying depression. However, persons with mental re-tardation may have symptom patterns more like thoseof children without mental retardation, because the twogroups tend to function in similar developmental ways(Charlot, Doucette, & Mezzacappa, 1993). For personswith more severe mental retardation, diagnosis is bet-ter when based on observations of behavior, sleep pat-terns, eating habits, and family patterns of the disorder.

Despite the devastating nature of depression, thediagnosis of depression is often missed in these indi-viduals. Rojahn, Warren, and Ohringer (1994) state thatthe most significant cause for the exclusion of personswith mental retardation from mental health services isthe difficulty of making a reliable and accurate diag-nosis of depression. Therefore the accurate assessmentof all aspects of depression, including its cognitive, af-fective, and behavioral manifestations, is crucial tohelping these individuals.

After depression is diagnosed, it can be treated.Pawlarcyzk and Beckwith (1987) suggest that conver-sations between a depressed client with mental retarda-tion and a therapist are beneficial in treating depression,as is social skills training. With modifications, treat-ments that have proven effective with the general pop-ulation should be useful in helping persons withintellectual limitations who are also depressed (e.g., an-tidepressant medication, light therapy for seasonal af-fective disorders, and behavior modification proceduressuch as pleasant-events scheduling, covert condition-ing, and functional behavioral analysis).

Anxiety is also a common psychological problemfor individuals with mental retardation (Matson,Smiroldo, Hamilton, & Baglio, 1997), but anxiety hasnot been as well researched as depression in this pop-ulation. Matson et al. (1997) report that prevalencerates of anxiety range from 2–25%. Ramirez andKratochwill (1997) found that children with mentalretardation, as compared to children without mentalretardation, were more likely to report specific fearsand generalized anxiety. However, there were moresimilarities than differences between these two groups.

As a construct, anxiety seems closely allied to de-pression. In persons without mental retardation, anxietyand depression measures are substantially correlated,with coefficients ranging from .50 to .80 (Kendall &Watson, 1989). This finding has been documented in avariety of populations, including normal adults, chil-dren, and psychiatric patients. Some investigators havesuggested that anxiety and depression cannot be easily

discriminated from each other, and, therefore, it is dif-ficult to determine which symptoms signal which dis-order (Steer, Beck, Riskind, & Brown, 1986). To date,this relationship has not been systematically studied inpersons with mental retardation.

Cognitive Theory of Depression and Anxiety

Considerable empirical support for the cognitivetheory of depression has been found. For example,Beck’s cognitive model relates depression to an unre-alistically negative view of the self, the person’s future,and the person’s world. Specifically, depressed indi-viduals’ cognitions center around themes of worthless-ness, incompetence, failure, and pessimism. Greenbergand Beck (1989) found that depressed participantsconsistently endorsed and recalled more negative andfewer positive depression-relevant stimuli than did non-depressed participants. The one study that examineddepressive symptoms and cognitions in adults with mildmental retardation found a strong positive relation bet-ween the two constructs (Nezu, Nezu, Rothenberg,DelliCarpini, & Groag, 1995).

As with depression, Beck believes that anxiety isa result of nonfunctional views held by the anxious per-son. However, he theorizes that anxious individuals’cognitions center around themes of threat, danger, un-predictability, and uncertainty, distinguishing them fromthe thoughts of depressed individuals, whose thinkingrevolves around personal worthlessness, incompetence,failure, and pessimism (Greenberg & Beck, 1989).

Cognitive Content-Specificity of Depression andAnxiety

Despite the aforementioned overlap of depressionand anxiety, most clinicians and researchers continueto conceptualize these emotions as separate constructssharing some common cognitive ground (Steer et al.,1986). In 1989, Greenberg and Beck predicted that de-pressed and anxious individuals could be differentiatedby specific cognitions. To test this hypothesis they useda trait-rating and incidental recall paradigm with 66participants. Although they obtained support for thecontent-specificity hypothesis of depression, supportfor the content-specificity hypothesis for anxiety wasweak. Research using the Cognitions Checklist (CCL)has resulted in a slightly different pattern of results(Beck, Brown, Steer, Eidelson, & Riskind, 1987). TheCCL was designed to assess the difference in cogni-tions between anxiety and depression and, therefore, toclarify the relationship between anxiety and depression.

It was found that the anxiety and depression subscalesof the CCL were consistent with the cognitive themesascribed to them.

The current study explored the relationship be-tween depression and anxiety in persons with mentalretardation. It was predicted that measures of depres-sion and anxiety would be correlated in persons withmental retardation, that the cognitive assessment ofmaladaptive thinking would predict depression andanxiety in this group, and that cognitive-specificitywould differentiate depression and anxiety.

METHOD

Participants

Participants were 46 people with a diagnosis ofmild (n 5 30), moderate (n 5 9), or borderline (n 5 7)mental retardation. Thirty-nine participants were Cau-casian, and seven were African Americans. The meanIQ of participants was 66.13 (SD 5 8.53, range 44–83). Forty-six percent of participants were female. Theaverage age of participants was 36.41 (SD 5 9.19,range 21–59). All participants were their own guardiansand were able to consent to participate. The study hadbeen approved by the Institutional Review Board andby the community agency.

Participants were recruited from a community-living and work program. The agency serves personswith mental retardation and allied developmental dis-abilities. All participants were living in supervisedapartments and were recommended by the director ofthe facility. From an original pool of 60, the directorrecommended 47 clients as being the most likely to beavailable for testing and the most willing to cooperate.None of the 47 refused to participate at first; however,during the study 1 participant dropped out because hebelieved the questions were too “personal.”

Records indicated that 7 participants were cur-rently diagnosed with a depressive disorder, although12 additional participants had formerly been diagnosedas depressed. (Consulting psychiatrists at a local men-tal health center had made diagnoses listed in the agencyrecords.) Three participants received antipsychoticmedication, 12 received antidepressants, 1 receivedantianxiety medication, and 6 received anticonvulsantmedication. (A diagnosis of depression was not a cri-terion for inclusion in this study; therefore, the highrate of depression may simply reflect the high co-occurrence of depression in this sample or a predilec-tion on the part of the evaluating psychiatrists todiagnosis depression in this group.)

Depression and Anxiety 71

Procedure

The senior author administered each of the scalesto participants. Although the senior author had workedpreviously with another unit of this community pro-gram, she had never worked with this unit or with anyof the participants, and she was blind to any diagnosesduring the testing phase.

Preliminary Questions

At the beginning of the session, four closed-endedquestions were administered to each participant to de-termine the possibility of a response bias (see Sigel-man, Budd, Spanhel, & Schoenrock, 1981). Each simplequestion (e.g., “Have you ever been to the moon?”)required a specific answer to be correct, either yes orno. Participants were required to answer all four ofthese questions correctly before the other testing mate-rials were administered. No participant missed any ofthese preliminary questions.

Test Administration and Chart Review

To ensure uniform administration, the senior au-thor then read all scale items to each participant. Thedecision to read the items was made by the senior au-thor, who was familiar with this program and concludedthat some of the clients could not read (e.g., one clienthad an IQ of 44).

To aid the respondent, the researcher also dis-played four flash cards with pictorial representationsof the possible answers, which were standardized forall four measures to include “never,” “a little,” “some-times,” and “a lot.” For the prompts, “a lot” was rep-resented by a card with 17 dots; “sometimes” by a cardwith 10 dots; “a little” by a card with 4 dots; and“never” by a blank card. The words representing thesechoices were also written on the cards. The order ofthese cards was reversed on a random basis to controlfor position bias. (All clients eventually became fa-miliar with the task and no longer depended on thecards.) The need to provide assistance to participantswith a visual prompt was based on the senior author’sknowledge of the cognitive limitations of some of themembers of this group. For example, it was believedthat some of the respondents would have initial diffi-culty maintaining, in active memory, the four levels ofthe rating scale. Also, the need for additional structureto reduce response bias had been suggested by the workof Sigelman et al., (Sigelman et al., 1981; Sigelman,Winer, & Schoenrock, 1982).

After the evaluations, the researcher completed afollow-up chart review of the participants, determining

72 Glenn, Bihm, and Lammers

for each participant any history of mental health andmedical problems, any psychological treatment, andcurrent functioning level.

Measures

The Beck Anxiety Inventory (BAI) (Beck & Steer,1990) is a 21-item self-report questionnaire that listssymptoms of anxiety. The BAI requires that the respon-dent rate how much each symptom has bothered him orher for the past week. It uses a 4-point scale ranging from“not at all” (0) to “severely” (3). A total score for the BAIis obtained by summing the ratings for the 21 items.

The Reynolds Child Depression Scale (RCDS)(Reynolds, 1989) is a 30-item self-report scale that askseach participant to rate his or her feelings for the past2 weeks on a 4-point scale ranging from “almost never”(1) to “all the time” (4). Items assess symptoms of de-pression (Reynolds & Graves, 1989). The RCDS waschosen because of the clarity of the item wordings.

The Automatic Thoughts Questionnaire (ATQ)(Hollon & Kendall, 1980) is a 30-item questionnairethat measures the frequency of automatic negative-thoughts associated with depression. Participants areasked how often they have certain thoughts, on a 5-pointrating scale from “not at all” (1) to “all the time” (5) andthe degree to which they believe these thoughts (on a5-point rating scale from “not at all” to “totally”). Eachof the scales is then summed, yielding two ATQ scores(frequency and strength). For the present study, onlythe frequency scale was administered, and this scalewas reduced to four choices, ranging from “never” to“a lot,” so that all tests used a 4-point scale. (The re-searchers deemed it important, in order not to confuseparticipants, to use the same rating scale, from 0 to 3,for all four measures.) The frequency ratings weresummed into two subscales measuring “Negative Self-

Concept and Negative Expectations” (NSC) and “Per-sonal Maladjustment and Desire for Change” (PM)(Joseph, 1994).

The Cognitions Checklist (CCL) (Beck et al.,1987) is a 25-item questionnaire designed to measurethe frequency of automatic thoughts relevant to anxietyand depression. The thoughts are rated on a 5-point scaleranging from “never” (0) to “always” (4). Like the ATQ,the CCL was modified for the present study by reduc-ing the number of scale choices from five to four. Thetwo subscales of the CCL, one measuring anxiety cog-nitions and the other measuring depressive cognitions,were included in the analysis of the present study.

RESULTS

The internal consistencies of the scales, as mea-sured by Cronbach alphas, were consistently high:RCDS, .92; BAI, .92; ATQ, .97; and CCL, .94. Becausethis is the first report using the BAI with persons withmental retardation, several additional analyses wereconducted. For gender, the mean scores, standard de-viations, and score ranges for the BAI for men andwomen were: 17.92, 14.41, 0–63, and 21.29, 14.92, 0–50, respectively. (Note: given that there were modifi-cations in the administration and scoring of all thescales, these group statistics should not be used asnormative data.) There were no statistical differencesbetween these scores. Likewise, there were no differ-ences based on age for the following groupings: below36 years and 36 and above. The means, standard devi-ations, and score ranges for the BAI for younger andolder participants were: 17.17, 12.17, 0–41, and 21.74,16.60, 0–63, respectively.

Table I reports the correlations among the totalscores and subscales of the four instruments adminis-

Table I. Correlation Matrix of Scales Measuring Anxiety, Depression, and Cognitions

Scale 1 2 3 3(a) 3(b) 4 4(a) 4(b)

1. RCDS –2. BAI .74 –3. ATQ .89 .80 –

a. NSC .86 .80 .98 –b. PM .87 .74 .96 .89 –

4. CCL .80 .76 .92 .92 .86 –a. Depression .76 .62 .88 .87 .83 .95 –b. Anxiety .73 .81 .85 .86 .76 .93 .77 –

Note.N 5 46. All correlations significant at p , .001. RCDS 5 Reynolds Child Depression Scale;BAI 5 Beck Anxiety Inventory; ATQ 5 Automatic Thoughts Questionnaire; NSC 5 Negative Self-Concept; PM 5 Personal Maladjustment; and CCL 5 Cognitions Checklist.

tered. All correlations were significant, and, as pre-dicted, anxiety and depression were highly correlated,r(44) 5 .74, p , .001.

We assessed the unique (nonoverlapping) contri-butions of specific cognitions to negative affect (as in-dicated by squared semipartial correlations). To do so,we conducted two hierarchical multiple regressionanalyses using SPSS (see Tables II and III). For bothanalyses (the first predicting depression as measuredby the RCDS and the second predicting anxiety as mea-sured by the BAI), the two subscales of the CCL wereentered as a block (model 1) and then the two subscalesof the ATQ were entered as the second block (model 2),as indicated in Table II.

Both models contributed to the prediction of de-pression (see Table II), explaining 64% and 79% of thevariance, respectively. For model 1, F(2, 43) 5 37.52,p , .001. For model 2, F(4, 41) 5 39.30, p , .001.Although both CCL subscales contributed unique vari-ance to depression in model 1, only the ATQ measuresof depression (and not the CCL measures of depressionand anxiety) uniquely predicted depression in the finalmodel.

A similar hierarchical regression examined jointand unique predictors (in terms of cognitions) of self-reported anxiety (see Table III). Once again, bothmodels predicted variance in anxiety, explaining 66%and 74% of the variance, respectively, model 1:F(2, 43) 5 42.17, p , .001, and model 2: F(4, 41) 528.46, p , .001. The CCL cognitive measure of anx-iety contributed unique variance to anxiety in bothmodel 1 and model 2, while the CCL cognitive mea-sure of depression was a negative predictor of anxi-ety in model 2.

Depression and Anxiety 73

DISCUSSION

Consistent with predictions based upon work withpersons without mental retardation (Beck et al., 1987;Hollon & Kendall, 1980; Steer et al.,1986), there wasa high correlation between anxiety and depression inthe current group and between negative affect and mal-adaptive cognitions. Furthermore, our results confirmthe preliminary findings of Nezu et al. (1995), whichsuggested a relationship between measures of depres-sion and cognitions in persons with mental retardation.The results also extend the cognitive model to the studyof anxiety in this population.

Both the RCDS and the BAI were easily adminis-tered to participants. A preliminary version of theRCDS has been evaluated on persons with mental re-tardation (Reynolds & Baker, 1988); however, to ourknowledge, ours was the first study to use the BAI withpersons with mental retardation. Using these measures,the relationship between depression and anxiety in thisgroup was quite high, which is not that surprising givensimilar findings from research with the general popu-lation (Kendall & Watson, 1989; Steer et al., 1986).These significant overlaps may have been due to ac-tual comorbidity or to problems in measuring the con-structs with the current scales. On the other hand, theshared variance between RCDS and the BAI was 55%,suggesting a less-than-perfect overlap. Thus, clini-cians and researchers should continue to assess bothconstructs.

Our findings also suggest that maladaptive cogni-tive processes can be assessed in persons with limitedintellectual abilities, and these cognitions can, in turn,be related to negative affect. Cognition accounted for

Table II. Hierarchical Multiple Regression Analysis PredictingDepression from Cognitive Variables on the Cognitions Checklist

(CCL) and Automatic Thoughts Questionnaire (ATQ)

Model and predictor variables R2 DR2 sr2 b

Model 1 .64 .64Depression (CCL) .10a .49a

Anxiety (CCL) .05a .36b

Model 2 .79 .16Depression (CCL) .00 2.05Anxiety (CCL) .00 2.03Negative Self-Concept (ATQ) .03a .48b

Personal Maladjustment (ATQ) .05a .50a

a p , .01.b p , .05.

Table III. Hierarchical Multiple Regression Analysis PredictingAnxiety from Cognitive Variables on the Cognitions Checklist

(CCL) and Automatic Thoughts Questionnaire (ATQ)

Model and predictor variables R2 DR2 sr2 b

Model 1 .66 .66Depression (CCL) .00 .00Anxiety (CCL) .27b .82a

Model 2 .74 .07Depression (CCL) .03c 2.38c

Anxiety (CCL) .07b .53b

Negative Self-Concept (ATQ) .02 .42Personal Maladjustment (ATQ) .01 .27

a p , .001.b p , .01.c p , .05.

79% of the variance of depression and 74% of the vari-ance in anxiety. Measures of anxiety-cognitions anddepressive-cognitions were also highly correlated (e.g.,there was a 59% overlap in variance between the de-pression and anxiety subscales of the CCL).

Consistent with the overall cognitive hypothesis,the ATQ and the CCL both predicted depression, ex-plaining 79% and 64%, respectively, of the variance ofself-reported depression (see Table I). Likewise, theATQ and the CCL both predicted anxiety, explaining64% and 58%, respectively, of the variance in self-re-ported anxiety (see Table I). Thus the cognitive hy-pothesis (that emotion can be predicted by cognition)was supported. However, there was scant support forthe cognitive-specificity hypothesis when analyzing thebivariate correlations (see Table I). Counter to the cog-nitive-specificity hypothesis, there was a high rela-tionship between depressive-cognition measures (i.e.,the two subscales of the ATQ and the depression sub-scale of the CCL) and the anxiety-cognition measure(the anxiety subscale of the CCL).

A strict cognitive-specificity hypothesis wouldposit that only cognitive measures of depression wouldpredict depression. However, as suggested by model 1in Table II, both the depression and anxiety cognitivesubscales of the CCL contribute to the variance of de-pression. However, when both depression and anxietysubscales of the CCL and ATQ are included in theequation (see model 2), the depressive cognitions mea-sured by the ATQ measures are the only unique con-tributors (as one might predict). The prediction ofanxiety uniquely by anxiety cognitions is more appar-ent in the results shown in Table III. In the final model,the CCL-anxiety cognitions predict anxiety, while theCCL depressive-cognition subscale is weighted nega-tively in predicting anxiety.

The lack of clear-cut support for cognitive-speci-ficity may be the result of several factors. First, giventhe great difficulty of developing “pure” measures ofanxiety or depression in any population, it may be dif-ficult to develop pure measures of their respective cog-nitions. There may be developmental limits to measuringboth components of negative affect. For example, youngchildren have undifferentiated emotions, at least as de-termined by self-report, and may lack the language tolabel emotions. They may also be unaware of the natureof internal processes, a precursor of accurate labeling(Schwartz, Goldstone, & Kaslow, 1998). In fact, a fre-quent observation of troubled children is that they are“not in touch with their feelings,” and they seem to op-erate with primitive theories about emotions and theircauses (Stipek & DeCotis, 1988). These cognitive lim-

74 Glenn, Bihm, and Lammers

itations may limit the assessment of these processes inpersons with mental retardation.

Still, knowing a client’s specific cognitions shouldhelp individualize treatment. In fact, an item-by-itemanalysis of both cognitive scales would be a good clin-ical approach. Using instruments such as these, a ther-apist could easily identify and address maladaptivethoughts in persons with mental retardation. Cliniciansand researchers would also want to assess the suitabil-ity of these individuals for cognitive therapy, perhapsby measuring their ability to identify emotions and tolink emotions to situations (Dagnan, Chadwick, &Proudlove, 2000).

Although our findings suggest that cognitive scalesdeveloped on nonretarded persons may be used to eval-uate persons with lower intellectual functioning, wewould urge researchers to consider developing cogni-tive scales that are specific to individuals with mentalretardation, because these persons may use unique de-scriptors and terms for emotional and cognitive expe-riences. These new scales could be developed frominterviews and from responses to projective assess-ments and could use the words, images, and analogiesprovided by persons with mental retardation. These in-struments could assess core beliefs, automatic thoughts,expectations, causal attributions, and ratings of self-efficacy. In addition, scales of suicide ideation andhopelessness need to be developed for persons withmental retardation. These measures could be related toother scales developed specifically for persons withmental retardation. Thus, valuable information wouldbe gained by relating them to instruments like the Di-agnostic Assessment for the Severely Handicapped, thePsychopathology Inventory for Mentally RetardedAdults, and the Aberrant Behavior Checklist (Aman &Singh, 1986; Bihm & Poindexter, 1991; Matson et al.,1984, 1991; Ramirez & Kratochwill, 1997).

Because significant modifications were made tothe instruments in the current study, it would not beappropriate to use the current results as normative datafor another sample. Given these modifications, it is dif-ficult to draw firm conclusions about the absolutedegree of anxiety or depression in the current sample.In the future, modifications to scales should not bemade unless absolutely necessary, for modificationchanges psychometric properties. The psychometricproperties of any revised scales would need to be care-fully documented.

In the current study, all participants might havebeen able to read at the required level, but this was notassessed. Future researchers might first assess readinglevel with a screening test and use the results as a cri-

terion for inclusion in the study (i.e., if it is requiredthat participants read the scales themselves, as was thecase in the original scale development). Other than in-ternal consistency, this study did not address issues ofvalidity and reliability (such as test-retest and self-otherreports). More attention to validity and reliability isessential for future scale development and use.

There are several other significant limitations ofthis study. The sample size was obviously small. “De-pression” was a current diagnosis for 15% of the cur-rent sample, but not one of the participants had a currentdiagnosis of an anxiety disorder (which is odd giventhe high correlation of anxiety and depression in thissample). It is possible the psychiatrists had underdiag-nosed anxiety. Because research of this nature shouldinclude persons with a diagnosable anxiety disorder,future participants might be evaluated with specificDSM-IV criteria (see Matson & Smiroldo, 1997) to en-sure the inclusion of participants with current anxietyand/or mood disorders.

Further, it is unclear how much our results wereaffected by the inclusion of seven participants diagnosedwith “borderline” intellectual functioning, a diagnosticcategory no longer included in the DSM system otherthan as a “v” code. Participants with the borderline di-agnosis did, however, have very low adaptive func-tioning, indicating a significant deficit in coping anddaily living skills. Therefore, we thought it appropriate,given the preliminary nature of this study, to includethis group. Finally, we recommend that researchers as-sess more diverse groups with mental retardation, ac-counting for the moderating effects of age, gender, andethnic and cultural factors, as they relate to cognitionand affect.

ACKNOWLEDGMENTS

This article was based upon Elizabeth Glenn’smaster’s thesis in the Department of Psychology andCounseling at the University of Central Arkansas. Theresearch was funded in part by a grant from the UCAUndergraduate Research Fund. We wish to thank thestaff and volunteers of Pathfinders, Inc., Jacksonville,AR, and two anonymous reviewers from this journal.

REFERENCES

Aman, M. G., & Singh, N. N. (1986). Aberrant Behavior Checklist:Manual.East Aurora, NY: Slosson Educational Publications.

American Psychiatric Association. (1994). Diagnostic and statis-tical manual of mental disorders(4th ed.). Washington, DC:Author.

Depression and Anxiety 75

Beck, A. T., Brown, G., Steer, R. A., Eidelson, J. I., & Riskind, J. H.(1987). Differentiating anxiety and depression: A test of the cog-nitive content-specificity hypothesis. Journal of Abnormal Psy-chology, 96,179–183.

Beck, A. T., & Steer, R. A. (1990). Beck Anxiety Inventory manual.San Antonio, TX: The Psychological Corporation.

Bihm, E. M., & Poindexter, A. R. (1991). Cross-validation of the fac-tor structure of the Aberrant Behavior Checklist for persons withmental retardation. American Journal on Mental Retardation,96, 209–211.

Charlot, L. R., Doucette, A. C., & Mezzacappa, E. (1993). Affectivesymptoms of institutionalized adults with mental retardation.American Journal on Mental Retardation, 98,408–416.

Dagnan, D., Chadwick, P., & Proudlove, J. (2000). Toward an as-sessment of suitability of people with mental retardation for cog-nitive therapy. Cognitive Therapy and Research, 24,627–636.

Greenberg, M. S., & Beck, A. T. (1989). Depression versus anxiety:A test of the content-specificity hypothesis. Journal of Abnor-mal Psychology, 98,9–13.

Hollon, S. D., & Kendall, P. C. (1980). Cognitive self-statements indepression: Development of an Automatic Thoughts Question-naire. Cognitive Therapy and Research, 4,383–395.

Joseph, S. (1994). Subscales of the Automatic Thoughts Question-naire. The Journal of Genetic Psychology,155, 367–368.

Kendall, P. C., & Watson, D. (Eds.) (1989). Anxiety and depression:Distinctive and overlapping features.San Diego, CA: Acade-mic Press, Inc.

Matson, J. L., Gardner, W. I., Coe, D. A., & Sovner, R. (1991). Ascale for evaluating emotional disorders in severely and pro-foundly mentally retarded persons: Development of the Diag-nostic Assessment for the Severely Handicapped (DASH) Scale.British Journal of Psychiatry, 159,404–409.

Matson, J. L., Kazdin, A. E., & Senatore, V. (1984). Psychometricproperties of the Psychopathology Instrument for Mentally Re-tarded Adults. Applied Research in Mental Retardation, 5,81–90.

Matson, J. L., Smiroldo, B. B., Hamilton, M., & Baglio, C. S. (1997).Do anxiety disorders exist in persons with severe and profoundmental retardation? Research in Developmental Disabilities, 18,39–44.

Matson, J. L., & Smiroldo, B. B. (1997). Validity of the mania sub-scale of the Diagnostic Assessment for the Severely Handi-capped-II (DASH-II). Research in Developmental Disabilities,18, 221–225.

Nezu, C. M., Nezu, A. M., Rothenberg, J. L., DelliCarpini, L., &Groag, I. (1995). Depression in adults with mild mental retar-dation: Are cognitive variables involved? Cognitive Therapy andResearch, 19,227–239.

Pawlarcyzk, D., & Beckwith, B. E. (1987). Depressive symptomsdisplayed by persons with mental retardation: A review. Men-tal Retardation, 25,325–330.

Ramirez, S. Z., & Kratochwill, T. R. (1997). Self-reported fears inchildren with and without mental retardation. Mental Retarda-tion, 35,83–92.

Reynolds, W. M. (1989). Reynolds Child Depression Scale: Profes-sional manual.Odessa, FL: Psychological Assessment Re-sources, Inc.

Reynolds, W. M., & Baker, J. A. (1988). Assessment of depressionin persons with mental retardation. American Journal on Men-tal Retardation, 93,93–103.

Reynolds, W. M., & Graves, A. (1989). Reliability of children’s re-ports of depressive symptomatology. Journal of Abnormal ChildPsychology, 17,647–655.

Rojahn, J., Warren, V. J., & Ohringer, S. (1994). A comparison ofassessment methods for depression in mental retardation. Jour-nal of Autism and Developmental Disorders, 24,305–313.

Schwartz, J. A. J., Goldstone, T. R. G., & Kaslow, N. J. (1998). De-pressive disorders. In T. H. Ollendick & M. Hersen (Eds.),Handbook of child psychopathology(3rd ed.) (pp. 269–289).New York: Plenum Press.

Sigelman, C. K., Budd, E. C., Spanhel, C. L., & Schoenrock, C. J.(1981). When in doubt, say yes: Acquiescence in interviewswith mentally retarded persons. Mental Retardation, 19,53–58.

Sigelman, C. K., Winer, J. L., & Schoenrock, C. J. (1982). The re-sponsiveness of mentally retarded persons to questions. Educa-tion and Training of the Mentally Retarded, 17,120–124.

Singh, N. N., Sood, A., Sonenklar, N., & Ellis, C. R. (1991). As-sessment and diagnosis of mental illness in persons with men-

76 Glenn, Bihm, and Lammers

tal retardation: Methods and measures. Behavior Modification,15, 419–443.

Steer, R. A., Beck, A. T., Riskind, J. H., & Brown, G. (1986). Dif-ferentiation of depressive disorders from generalized anxiety bythe Beck Depression Inventory. Journal of Clinical Psychology,42, 475–478.

Stipek, D. J., & DeCotis, K. M. (1988). Children’s understanding ofthe implications of causal attributions for emotional experiences.Child Development, 59,1601–1616.