IQ and Scores on the Mini-Mental State Examination (MMSE): Controlling for Effort and Education Among Geriatric Inpatients

ABSTRACT

The Mini-Mental State Examination (MMSE) and the Peabody Picture Vocabulary Test, Third Edition (PPVT-III) were administered to 76 Extended Care Center inpatients of a Veterans Affairs Medical Center, 56 of whom did not carry a diagnosis associated with brain impairment, and 20 of whom had a diagnosis associated with degenerative or cerebrovascular conditions. MMSE and PPVT-III scores, the latter of which estimates verbal IQ, were significantly correlated in both groups, even when controlling for the effects of education. This correlation was attenuated, but remained significant, in the face of varying levels of patient effort. The findings of this study support earlier findings, which also suggest that interpretation of “impaired” performance on the MMSE for those with IQ scores below 90 should be interpreted with caution. Average MMSE scores for Wechsler IQ classifications are provided.     

IQ and scores on the Mini-Mental State Examination (MMSE): controlling for effort and education among geriatric inpatients

The Mini-Mental State Examination (MMSE) and the Peabody Picture Vocabulary Test, Third Edition (PPVT-III) were administered to 76 Extended Care Center inpatients of a Veterans Affairs Medical Center, 56 of whom did not carry a diagnosis associated with brain impairment, and 20 of whom had a diagnosis associated with degenerative or cerebrovascular conditions. MMSE and PPVT-III scores, the latter of which estimates verbal IQ, were significantly correlated in both groups, even when controlling for the effects of education. This correlation was attenuated, but remained significant, in the face of varying levels of patient effort. The findings of this study support earlier findings, which also suggest that interpretation of "impaired" performance on the MMSE for those with IQ scores below 90 should be interpreted with caution. Average MMSE scores for Wechsler IQ classifications are provided.     

Psychometric properties of the Folstein Mini-Mental State Examination

Criterion-referenced (Livingston) and norm-referenced (Gilmer-Feldt) techniques were used to measure the internal consistency reliability of Folstein's Mini-Mental State Examination (MMSE) on a large sample (N = 418) of elderly medical patients. Two administration and scoring variants of the MMSE Attention and Calculation section (Serial 7s only and WORLD only) were investigated. Livingston reliability coefficients (rs) were calculated for a wide range of cutoff scores. As necessary for the calculation of the Gilmer-Feldt r, a factor analysis showed that the MMSE measures three cognitive domains. Livingston's r for the most widely used MMSE cutoff score of 24 was .803 for Serial 7s and .795 for WORLD. The Gilmer-Feldt internal consistency reliability coefficient was .764 for Serial 7s and .747 for WORLD. Item analysis showed that nearly all of the MMSE items were good discriminators, but 12 were too easy. True score confidence intervals should be applied when interpreting MMSE test scores.     

Correlates of level and change in the Mini-Mental State Examination

The goal of the current project was to determine (a) the cognitive abilities assessed by the Mini Mental State Examination (MMSE; M. F. Folstein, S. E. Folstein, & P. R. McHugh, 1975), (b) whether the same abilities are associated with MMSE performance among people of different ages, and (c) whether the same abilities are involved in changes within the same people over time. The authors therefore examined whether the initial levels of 5 cognitive ability constructs (vocabulary, reasoning, memory, space, and speed) predicted initial levels of MMSE performance and whether the initial levels or the changes in these 5 cognitive abilities predicted change in the MMSE performance. The major findings were that 3 cognitive constructs (vocabulary, reasoning, and memory) contribute to performance in the MMSE but that their respective contributions to the MMSE vary as a function of age and time. In particular, individual differences and change in the MMSE are primarily related to individual differences in reasoning among adults younger than about 70 years, whereas both initial level of MMSE performance and longitudinal change in MMSE performance primarily related to initial level and change in memory ability among adults older than 70 years. The results therefore imply that both the level of performance on the MMSE at a single point in time and the change in MMSE over time may represent somewhat different cognitive abilities at different ages.     

Neuropsychometric correlates of the Mini-Mental State Examination: preliminary data

The present study investigated the neuropsychometric correlates of the Mini-Mental State Examination. 12 consecutive neurological referrals were administered the Mini-Mental State Examination, the Wechsler Memory Scale, and the WAIS--R. Pearson product-moment correlations suggest moderate association between scores on the Mini-Mental State Examination and Wechsler Memory Scale but less robust relationships between scores on the Mini-Mental State Examination and WAIS--R Full Scale, Verbal, and Performance IQs.     

Factor structure of the Mini-Mental State Examination in adult psychiatric inpatients

The factor structure of the Mini-Mental State Examination (MMSE) has been examined primarily in geriatric populations using differing methodology and yielding inconsistent results. This study examined the MMSE in 339 adult inpatients at a nonforensic state psychiatric hospital. Data were analyzed in the form of 30 dichotomous variables (1 for each scored response). A 5-factor oblique solution was retained, comprising factors of Orientation, Attention-Working Memory, Verbal Recall, Comprehension-Praxis, and Naming. Higher order factors of Executive Functioning and Language-Praxis were also found. The results fit the traditional neurocognitive domains of a mental status examination and generally replicated findings from the 2 largest studies published to date despite differences in population, suggesting that the obtained factors are fairly robust.     

Structure of the Mini-Mental State Examination: a comparison between normal and clinical subjects

The Mini-Mental State Examination was administered to a sample of 133 Italians ranging from 60 to 80 years of age. From a Principal Components Analysis three factors were retained, being results of both varimax and oblimin rotations. These were labeled (1) Memory, (2) Verbal Comprehension, and (3) Attention. These factors are similar to those obtained with normal elderly Americans, but they are different from the factors based on clinical samples. No age differences in the factor scores were found among the normal elderly subjects when education was considered. In contrast, significant differences were found between a group of 25 demented patients and a matched group randomly extracted from the normal sample.     

Sex Differences in Self-Estimates on Two Validated IQ Test Subscale Scores

Two related studies investigated sex differences in self- and parental estimates of IQ scores on specific scales derived from standardized and validated IQ tests. In the first study 210 participants were asked to estimate their scores on the 11 WAIS-R subtests as well as their overall general IQ. Results showed males estimated their overall score, plus their total WAIS score, significantly higher than females, with effect sizes around 0.5. Factor analysis showed participants did differentiate between verbal and performance subscale scores on this test. In the second study 117 participants performed a similar task, but this time on the 12 subscales of the Stanford-Binet test. Males estimated their overall score higher than females. Factor analysis also showed 2 clear factors that reflected exactly the verbal and performance subscale scores.     

Seemingly Anomalous WISC-IV Full Scale IQ Scores in the American and Canadian Standardization Samples

The American and Canadian standardization samples for the fourth edition of the Wechsler Intelligence Scale for Children were used to ascertain the prevalence of “seemingly anomalous” Full Scale IQ (FSIQ) scores for all possible sums of scaled scores combinations of the Verbal Comprehension Index, Perceptual Reasoning Index, Working Memory Index, and Processing Speed Index. About 1% of FSIQ scores in both samples were anomalous (that is, fell outside the range of the four indices). Interestingly, the prevalence of anomalous scores was higher at the extremes of the FSIQ distribution. Moreover, anomalous FSIQ scores occurring above or below the mean of the FSIQ distribution (that is, 100) were larger or smaller, respectively, than any of the four indices. Furthermore, if the range of the four indices did contain the FSIQ (the non-anomalous case), when the FSIQ was below or above 100 it tended to be located distally in the half-segment of the range furthest below or above 100, respectively. A non-technical intuitive analogy from Olympic sports and a statistical explanation for the findings are provided. Important implications for practitioners are also presented. Portions of this paper were presented March 17, 2006 at the 20th annual Joseph R. Royce Research Conference, Department of Psychology, University of Alberta, Edmonton, Alberta.     

Seemingly Anomalous WISC-IV Full Scale IQ Scores in the American and Canadian Standardization Samples

The American and Canadian standardization samples for the fourth edition of the Wechsler Intelligence Scale for Children were used to ascertain the prevalence of “seemingly anomalous” Full Scale IQ (FSIQ) scores for all possible sums of scaled scores combinations of the Verbal Comprehension Index, Perceptual Reasoning Index, Working Memory Index, and Processing Speed Index. About 1% of FSIQ scores in both samples were anomalous (that is, fell outside the range of the four indices). Interestingly, the prevalence of anomalous scores was higher at the extremes of the FSIQ distribution. Moreover, anomalous FSIQ scores occurring above or below the mean of the FSIQ distribution (that is, 100) were larger or smaller, respectively, than any of the four indices. Furthermore, if the range of the four indices did contain the FSIQ (the non-anomalous case), when the FSIQ was below or above 100 it tended to be located distally in the half-segment of the range furthest below or above 100, respectively. A non-technical intuitive analogy from Olympic sports and a statistical explanation for the findings are provided. Important implications for practitioners are also presented. Portions of this paper were presented March 17, 2006 at the 20th annual Joseph R. Royce Research Conference, Department of Psychology, University of Alberta, Edmonton, Alberta.     

State IQ and suicide rates in the United States

Previous findings pertaining to the direction (positive or negative) of the ecological correlation between intelligence and suicide rates in the USA have been conflicting. Using novel state IQ estimates, derived from the Scholastic Assessment Test, the American College Test, these tests combined, or the National Assessment of Educational Progress state scores, these estimates were not consistently associated to state suicide rates. Whereas ACT-derived state IQ was significantly positively correlated with suicide rates, the correlation with composite ACT-SAT-derived state IQ was significantly negative and with both SAT-derived and NAEP-derived state IQ also negative but not significant. Validity checks pointed to possible methodological problems with the state IQ estimates. Currently available estimates of state IQ, therefore, seem not appropriate to resolve the question of the direction of the ecological correlation of intelligence and suicide mortality across the USA.     

State IQ and elderly suicide rates in the United States

Paralleling previous findings with state suicide rates of the total population, the associations of state suicide rates of elderly persons with regional IQ estimates across the USA were inconsistent (positive, negative, or nil), depending on the source of available state IQ estimates used in the analysis. The implications of these findings and directions for further inquiry are discussed.     

State preferences for the ACT versus SAT complicates inferences about SAT-derived state IQ estimates: A comment on Kanazawa (2006)

Kanazawa [Kanazawa, S. (2006). IQ and the wealth of states. Intelligence, 34, 593–600.] offered estimates of state IQ derived from SAT data. The purpose of this commentary is to argue that state preferences for the use of the ACT versus the SAT create biased estimates of SAT-derived state IQ for states where the ACT is more frequently used than the SAT. This error can be reduced by using both ACT and SAT data to estimate state IQ. An IQ estimate based on a ACT-SAT composite and a NAEP-derived state IQ estimate were compared as predictors of three wealth variables. Both IQ estimates cause one to conclude that states with higher mean IQ have larger gross state product per capita, higher median incomes, and a lower percentage of their population in poverty.     

Effects on WAIS-III performance IQ (PIQ) and full scale IQ (FSIQ) when object assembly is substituted for each standard performance scale subtest

According to the WAIS-III Administration and Scoring Manual, Object Assembly (OA) may be substituted for any spoiled Performance subtest. This assertion has not been evaluated in a clinical sample. The present investigation reports differences that resulted in Performance IQ (PIQ) and Full Scale IQ (FSIQ) when OA replaced each of the Performance subtests. Participants were 47 referrals for neuropsychological assessment (age M = 45.98 years, SD = 9.82; education M = 13.82 years, SD = 2.78). Results indicated that OA may replace any Performance subtest without seriously altering the summary scores. Differences between the standard IQs and OA-based composites were < 2 points for PIQ and < 1 point for FSIQ. More than 90% of the OA-based composites fell within the 90% confidence limits of the corresponding IQ.     

Meta‐Analyses of the Relationship of Creative Achievement to Both IQ and Divergent Thinking Test Scores

There is disagreement among researchers about whether IQ tests or divergent thinking (DT) tests are better predictors of creative achievement. Resolving this dispute is complicated by the fact that some research has shown a relationship between IQ and DT test scores (e.g., Runco & Albert, 1986; Wallach, 1970). The present study conducted meta‐analyses of the relationships between creative achievement and both IQ and DT test scores. The analyses included 17 studies (with 5,544 participants) that established the correlation coefficients between IQ and creative achievement and 27 studies (with 47,197 participants) that established the correlation coefficients between DT test scores and creative achievement. Marginal, but statistically significant, Fisher's Z‐transformed correlation coefficients were revealed. The analysis found a significantly higher relationship between DT test scores and creative achievement (r = .216) than between IQ test scores and creative achievement (r = .167). The differences in the correlation coefficients were explained by differences in DT tests, creative achievement types, predicted time periods, and creativity subscales. The significant independent moderator effect for different DT tests indicates that the Torrance Tests of Creative Thinking (TTCT) predict creative achievement better than any other DT test included in this study. Among the creative achievement types, music is predicted the best by IQ and all others are predicted best by DT tests. Among the time periods evaluated, the relationship between DT test scores and creative achievement had the highest correlation at the period of 11–15 years.     

An Examination of the Selection of Criterion Scores for the Study of Motor Learning and Retention

150 Ss were given 50 practice trials on the stabilometer over a period of 4 days (20 trials on Day 1 followed by a 1-day rest, 10 trials on Day 2, followed by a 7-day rest, 10 trials on Day 3 followed by a 14-day rest, and 10 trials on Day 4). A series of learning scores reflecting improvement in performance over the 4 days and retention scores reflecting loss of performance between days were obtained. The reliability of individual differences in these scores were calculated using the odd-even method and by progressively increasing the number of trials included in the measure of initial and final performance level. On the basis of the findings obtained it was concluded that using 4 – 6 trials in initial and final ability levels results in a difference (criterion) score which combines relatively high reliability with a relatively high estimate of the amount of change in performance.     

Holtzman Inkblot Anxiety and Shading Scores Related to State and Trait Anxiety

The State-Trait Anxiety Inventory (STAI) and the group Holtzman Inkblot Technique (HIT) were administered to 52 college Ss in order to determine the relationship between HIT Anxiety (Ax) and Shading (Sh) scores and A-Trait and A-State scores. Neither Ax or Sh correlated with A-Trait. Ax correlated with A-State, both before and during the administration of the inkblots. To determine the influence of productivity or inhibition, inkblot protocol word counts were obtained. Word count failed to correlate with the sc ores on the STAI, but did correlate significantly with the HIT scores.     

Meta-Analysis of Intelligence Quotient (IQ) in Obsessive-Compulsive Disorder

Obsessive compulsive disorder (OCD) is associated with a moderate degree of underperformance on cognitive tests, including deficient processing speed. However, despite little research focusing on Intelligence Quotient (IQ) in OCD, it has long been speculated that the disorder is associated with elevated intellectual capacity. The present meta-analytic study was, therefore, conducted to quantitatively summarize the literature on IQ in OCD systematically. We identified 98 studies containing IQ data among individuals with OCD and non-psychiatric comparison groups, and computed 108 effect sizes for Verbal IQ (VIQ, n = 55), Performance IQ (PIQ, n = 13), and Full Scale IQ (FSIQ, n = 40). Across studies, small effect sizes were found for FSIQ and VIQ, and a moderate effect size for PIQ, exemplifying reduced IQ in OCD. However, mean IQ scores across OCD samples were in the normative range. Moderator analyses revealed no significant moderating effect across clinical and demographic indices. We conclude that, although lower than controls, OCD is associated with normative FSIQ and VIQ, and relatively lowered PIQ. These results are discussed in light of neuropsychological research in OCD, and particularly the putative impact of reduced processing speed in this population. Recommendations for utilization of IQ tests in OCD, and directions for future studies are offered.