Table 5.

Predicted means and marginal effects for lexical ability from the linear regression models with interaction effects based on the refined model (6 items), by subgroup and year.

Note: Estimates come from lm(Theta6 ~ year + sex + age_group + sel_group + year:age_group + sex:age_group). Predicted means represent the model-based expected values of lexical ability for each subgroup and year, holding other variables constant. Marginal effects indicate the estimated change in lexical ability associated with a one-unit difference in the variable of interest, averaged across the sample.

Table 6.

Predicted means and marginal effects for lexical ability from the linear regression models with interaction effects based on the original 20 items, by subgroup and year.

Note: Estimates come from lm(Theta20 ~ year + sex + age_group + sel_group + year:age_group + sex:age_group). Predicted means represent the model-based expected values of lexical ability for each subgroup and year, holding other variables constant. Marginal effects indicate the estimated change in lexical ability associated with a one-unit difference in the variable of interest, averaged across the sample.

As shown in the Bland-Altman plot (Figure 2), some test takers have their lexical abilities overestimated or underestimated when the original 20 items are used compared to the six-item version with improved measurement properties. While most test takers’ lexical ability was not significantly over- or underestimated, some differences were observed across gender and years. Chi-square test confirmed a statistically significant but negligible association between sex χ²(2) = 27.03, p < .001, Cramér’s V = 0.03 and years χ²(4) = 57.79, p < .001, Cramér’s V = 0.03. Males were slightly more likely to have underestimated lexical ability with the original 20 items, whereas females were more likely to have overestimated lexical ability with the original 20 items. Pairwise tests revealed that the proportion of test takers underestimated with the original 20 items was slightly higher in 2022 compared to 2014 and 2018, while differences in overestimation were small but significant in 2022 compared to earlier years.

Figure 2. Bland-Altman plot comparing estimated lexical ability based on the refined model with six items and the original 20 items.

4. Discussion

This study aimed to enhance current knowledge of changes in measurement properties may influence results from intelligence tests. By using data from a synonyms test from the Swedish SEB-2000 test, used for selection to basic military training, as an example, the main findings in this paper display fit properties and invariance issues in the synonym test. Yet it shows that the measurement properties could be improved by removing malfunctioning items. While the refined item composition improved measurement properties, except for reliability, it had only a minor impact at the group level when comparing lexical ability over time and across subgroups of test takers for admission tests to basic military training in Sweden.

The major issues of misfit and invariance in the synonym test studied here could be addressed by refining item composition, though this approach yielded lower reliability and a risk of construct underrepresentation. The implications of improved measurement properties for group-level comparisons were small, but it should be noted that individuals at the upper and lower ends are more affected than those in the middle. Specifically, the lexical ability of some test takers at the lower end of the scale was likely overestimated, whereas it was likely underestimated at the upper end of the scale if all items were used, including those not showing satisfactory measurement properties. This pattern is expected, as reliability and precision are lower at the extremes of the scale, which explains why the observed differences occur at the ends of the scale. Currently, when test takers are selected for military training and various positions, the information from SEB (i.e., not only a single synonym test as used here) is primarily based on cut-off scores that fall in the middle or slightly above or below. Thus, differences between upper and lower-end individuals may not have any substantial influence on selections today. Nevertheless, such effects may be important in the future, for instance, if intelligence tests are used to match test takers to different positions or if selection thresholds must be lowered to further reinforce the Swedish Armed Forces. While it is also likely that similar problems with misfitting items and invariance issues may arise in other testing contexts, the magnitude and practical implications of such effects can naturally vary across settings and types of decision-making. Furthermore, regarding the effects of improved measurement properties when adding multiple subtests together to provide a measure of GMA, it is recommended to conduct further studies, for instance, by so-called differential test functioning (Stark et al., 2004), focusing on measurement properties at the test level (e.g., when results from subtests are combined to a global estimate).

Our results corroborate previous work, such as on the ASVAB and NIT, showing item drift across multiple verbal items (Chan et al., 1999; Shiu et al., 2013). Importantly, although the other studies covered longer periods, 16 and 72 years, respectively, this study only spanned eight years, while still indicating risks of item drift. Moreover, as discussed in the study of the Norwegian Armed Forces’ word similarities test, certain vocabulary and concepts may have been more familiar and relevant at the time of the test’s development, making them more challenging for newer generations (Nordmo et al., 2025). However, why certain words are easier or more difficult over time and in different groups remains to be explained, for instance, by constructing specification equations inspired by related word-based tests (Stenner et al., 2006; Melin et al., 2022; Bulut et al., 2023; Svetina et al., 2011) or through qualitative approaches that explore contextual and cultural factors influencing word familiarity. This can enable better designs of verbal items and guide how often verbal items need to be re-examined or replaced in test batteries.

The composition of subgroup characteristics differed across the annual cohorts in the analyses. The composition and representativeness of 17-19-year-old men in the admission testing were rather stable until 2006, but since then, with periods of compulsory or non-compulsory admission tests for basic military training and stressed gender equality in the Swedish Armed Forces, the estimated coverage has changed (Ludvigsson et al., 2022). Thus, the present study did not aim to evaluate trends in lexical ability but rather to examine the consequences of measurement properties when evaluating such trends. However, it is worth noticing a major difference in lexical ability between 2014 and 2018, when compulsory admission tests were reintroduced in 2017. Whether this is evidence for a general change in lexical ability or reflects a changed composition of subgroups remains to be further evaluated, preferably with a similar study design and including additional subtests of other intelligence components. Moreover, it should be noted that variations in subgroup composition, as demonstrated in this study, are likely to occur in other contexts as well, highlighting the necessity of carefully considering both the demographic structure under investigation and the broader societal or contextual factors that may exert influence.

Females generally perform better than males in the Swedish language subject in school (Swedish Association of Local Authorities and Regions, 2019; Swedish National Agency for Education, 2023) and verbal cognitive test (see summary in Roebuck-Spencer et al., 2008), yet males perform better on the synonym test here. Moreover, as reported by McGeown et al. (2016), females read more frequently, have more positive attitudes toward reading, higher reading motivation, greater confidence in their reading skills, and superior reading abilities than males. The fact that males perform better can be a sampling issue in the present study, possibly due to a first selection based on the mustering questionnaire (an online web survey including biodata questions). Nevertheless, it may also be related to, and questioning, the synonym test's validity. Could the synonyms used in the SEB-2000 test be "male-attributed" words, thereby introducing bias and disadvantaging female test takers? Lexical ability, as measured by the synonym test, is related to reading ability, which is strongly shaped by motivation and interest in texts. For instance, in elementary school, females generally read more stories about relationships and romance, whereas males generally consume more action-packed content of science fiction or sports featuring male protagonists (c.f. Clark & Foster, 2005; Lepper et al., 2022; McGeown et al., 2016). However, at this stage, we have not investigated whether the synonyms used in the SEB-2000 test could be considered as typically "male-female-attributed" words. Thus, with possible new synonym tests and with the proportion of females increasing in the admission testing, deeper investigations are required to avoid biased disqualification of certain groups in military selection.

The Flynn Effect, that is, that more recent cohorts show higher test scores in the intelligence domains (Flynn, 2012; Williams, 2013), is well known. On the other hand, more recent studies have also presented a negative Flynn Effect (Dutton & Lynn, 2013; Meisenberg & Lynn, 2023), and it must also be noted that intelligence trends are too rich and diverse to be captured by any single construct (Flynn, 2012). Previously, Swedish musters between 1970 and 1993 showed greater improvement over the years in test takers’ visual-spatial abilities, but little change and a slight decline in crystalized intelligence (Rönnlund et al., 2013). Others have proposed differences in trends across intelligence domains, which could be explained by societal cognitive development (Alwin & Pacheco, 2012) and a shift toward greater emphasis on rational thinking and analytic abilities (Flynn, 2012). Suppose the decrease in lexical ability presented here reflects a negative Flynn Effect. In that case, differences in one of many intelligence components or changes in group composition remain to be further studied. Importantly, a recent work on this topic on the Norwegian Armed Forces concludes that: previous studies using the Norwegian GMA data must be interpreted with more caution, but that the test should measure males and females equally fairly (Helland-Riise et al., 2024). This is of critical importance, likely also applicable in Sweden and other countries, and further work is warranted to ensure efficient and fair testing.

Methodological Considerations

First, it is worth noticing that many of the 20 synonym items studied here, until recently used for admission tests for basic military training in Sweden, do not fit the Rasch model. To our knowledge, no Rasch analysis has been conducted to evaluate measurement properties when implementing SEB to date. Thus, we cannot know when such misfitting issues have occurred or ascertain the reasons why so many items show poor fit to the measurement model. Furthermore, it must be noted that not even the refined model with six items fits the Rasch model perfectly, although a too-hardline statistical approach is not always desirable (Morel & Cano, 2017). Moreover, at the earlier analytical stages, alternative approaches could have included splitting items with DIF and forming testlets for locally dependent items, thereby retaining items and improving reliability. In the present case, where we sought to shed light on the importance of continuous evaluation of measurement properties – in particular examinations of possible item drift – and its implications, the refined model was considered satisfactory for our illustrative purposes.

As stated above, the removal of more than two-thirds of the items in the refined item set raises concerns about potential construct underrepresentation, that is, the extent to which a test fails to comprehensively capture the intended construct (Joint Committee on the Standards for Educational and Psychological Testing, 2014). It is important to emphasize that this study does not propose simply omitting the 14 excluded items in today’s testing. Rather, it forms part of the SDCAA’s ongoing efforts to modernize and improve the SEB. Drawing on historical data, our findings highlight the need for continuous test evaluation to ensure fairness and validity in selection procedures. However, decisions regarding how SDCAA should replace malfunctioning items or subtests lie beyond the scope of this study. Moreover, while item removal may introduce risks of construct underrepresentation, it may simultaneously enhance construct-irrelevant variance (Messick, 1995), particularly when outdated items are removed. For example, a word such as pandemic—which was rarely used in everyday language before 2020—may now be easily recognized by most test takers, not because of improved lexical ability, but because of increased societal exposure. Over time, certain words may become more or less relevant to the intended construct and may therefore warrant removal or replacement due to construct relevance. Verbal ability tests must account for such linguistic and cultural shifts in order to preserve construct validity and minimize the introduction of irrelevant variance arising from changing patterns of language familiarity.

A close inspection of the CICC plots becomes important with a high misfit. For instance, SYNONYMS 11 and 17 showed a particularly high misfit and a significant deviation from the curve, with dots flatter than expected. This implies that the item does not discriminate enough and can be a sign of guessing (Andrich et al., 2012). In contrast to other IRT models, such as the 3PL model, the Rasch model does not include a separate guessing parameter. Notably, in the 3PL model, guessing is attributed to the item. On the other hand, it is persons and not items that guess (Andrich & Marais, 2019), and therefore, guessing is more likely to be considered an interaction between the propensity of an item to provoke guessing and the proclivity of a person to guess (Smith, 1993). While potential guessing could be explored further, we decided to remove those items from the refined model in our study.

In contrast to the work on the entire ASVAB (Chan et al., 1999) and the Estonian version of NIT (Shiu et al., 2013), this paper examined item drift only in one subtest in the SEB-2000. However, this paper also introduced the role of using Rasch analysis for including longitudinal studies of measurement properties of lexical ability intelligence tests to provide an enhanced understanding of how measurement properties may have changed over time and their implications for understanding trends in lexical ability. Nevertheless, we welcome similar studies focusing on the other subtests in the SEB-2000 and the newly launched SEB-2025, as well as on additional test batteries, in future research. We also invite further research incorporating additional data to evaluate the test’s predictive validity and practical impact in selection contexts. Such efforts could contribute to the continued refinement of its measurement properties over time, as well as to longitudinal studies examining other domains of intelligence and the overall selection process.

5. Conclusion

This study aimed to enhance current knowledge of how changes in measurement properties may influence results from intelligence tests. Using data from the Swedish SEB-2000, the study identified fit and invariance issues in the synonym test, yet it also shows that the measurement properties could be improved by removing malfunctioning items. After identifying misfit and invariance issues, removing malfunctioning items improved the measurement properties of a synonym test used in the SEB-2000, but also lowered its reliability. While the implications of improved measurement properties for group-level comparisons were small, it must be noted that individuals at the upper and lower ends are more affected than those in the middle, which may threaten fairness in the selection process. Moreover, it is likely that similar problems with misfitting items and invariance issues may arise in other testing contexts, although the magnitude and practical implications of such effects can naturally vary across settings and types of decision-making.

Citing Chan et al. (1999, p. 617): “Test development and revision is an ongoing process. Test developers and users should constantly evaluate the psychometrics of their items and tests to guard against the possibility of DIF over time”, which is corroborated in this study. Importantly, the SDCAA has initiated such work to refine the existing SEB by replacing malfunctioning items and subtests to ensure high quality and fairness in their decision-making for admission to basic military training in Sweden. It is also likely to be needed in other countries and authorities that use admission tests over time. Continuous revision of admissions tests would not only benefit the individual organization by making test administration fairer and more efficient, but also enable the sharing of results, creating opportunities to design and trustfully manage tests better.

Given the evolving composition of test takers for basic military training in Sweden, including changes in sex, age, and selection groups, ongoing evaluations like this kind of study are essential to maintain fair selections. Furthermore, this study underscores the need for regular reassessments of measurement properties – and actions to ensure invariance –to better understand how intelligence may change over time. The results in this study primarily pertain to lexical ability among test takers for admission to basic military training in Sweden. However, the broader implications for test validity and fairness are relevant across many fields where standardized testing plays a critical role.

Funding and Conflict of Interest

This study is a part of the Swedish Defence Conscription and Assessment Agency's work to update the existing Swedish Enlistment Battery. The work has been partially funded by the Swedish Defence Conscription and Assessment Agency. Beyond that, the authors have no conflicts of interest to declare.

Data Availability Statement

The data that support the findings of this study have been retrieved from the Swedish Defence Conscription and Assessment Agency. Restrictions apply to the availability of these data, which were used under license for this study.

How to Cite

Melin, J., Wahlkrantz, E., Brandebo, M. F., Larsson, G., & Annell, S. (2026). Are intelligence measures stable across time and groups? A study of invariance in a synonyms test. Educational Methods & Psychometrics, 4: 22. https://dx.doi.org/10.61882/emp.2026.1

Allaire, J. J., Teague, C., Scheidegger, C., Xie, Y., & Dervieux, C. (2022). Quarto (Version 1.2) [Computer software]. https://doi.org/10.5281/zenodo.5960048

Alwin, D. F., & Pacheco, J. (2012). Population trends in verbal intelligence in the United States. In 13. Population Trends in Verbal Intelligence in the United States (pp. 338–368). Princeton University Press. https://doi.org/10.1515/9781400845569-015

Andrich, D. (1982). An index of person separation in latent trait theory, the traditional KR-20 index, and the Guttman scale response pattern. Education Research and Perspectives, 9(1), 95–104. https://doi.org/10.3316/aeipt.13636

Andrich, D. (2004). Controversy and the Rasch Model: A characteristic of incompatible paradigms? Medical Care, 42(Supplement), I–7. https://doi.org/10.1097/01.mlr.0000103528.48582.7c

Andrich, D., & Hagquist, C. (2012). Real and artificial differential item functioning. Journal of Educational and Behavioral Statistics, 37(3), 387–416. https://doi.org/10.3102/1076998611411913

Andrich, D., & Marais, I. (2019). A Course in Rasch measurement theory: Measuring in the educational, social and health sciences. Springer Singapore. https://doi.org/10.1007/978-981-13-7496-8

Andrich, D., Marais, I., & Humphry, S. (2012). Using a theorem by Andersen and the dichotomous Rasch model to assess the presence of random guessing in multiple choice Items. Journal of Educational and Behavioral Statistics, 37(3), 417–442.

Arel-Bundock [aut, V., cre, cph, Greifer, N., Bacher, E., McDermott, G., & Heiss, A. (2025). marginaleffects: Predictions, comparisons, slopes, marginal means, and hypothesis tests (Version 0.30.0) [R package]. https://cran.r-project.org/web/packages/marginaleffects/index.html

Babcock, B., & Albano, A. D. (2012). Rasch scale stability in the presence of item parameter and trait drift. Applied Psychological Measurement, 36(7), 565–580. https://doi.org/10.1177/0146621612455090

Bäckström, P. (2022). Self-selection and recruit quality in Sweden’s all volunteer force: Do civilian opportunities matter? Defence and Peace Economics, 33(4), 438–453. https://doi.org/10.1080/10242694.2021.1903284

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67, 1–48. https://doi.org/10.18637/jss.v067.i01

Bland, J. M., & Altman, D. G. (1999). Measuring agreement in method comparison studies. Statistical Methods in Medical Research, 8(2), 135–160. https://doi.org/10.1177/096228029900800204

Bond, T., Yan, Z., & Heene, M. (2020). Applying the Rasch model: Fundamental measurement in the human sciences (4th ed.). Routledge.

Borsboom, D. (2006). When does measurement invariance matter?: Medical Care, 44(Suppl 3), S176–S181. https://doi.org/10.1097/01.mlr.0000245143.08679.cc

Buchardt, A.-S., Christensen, K. B., & Jensen, N. (2023). Visualizing Rasch item fit using conditional item characteristic curves in R. Psychological Test and Assessment Modeling, 65(2), 206–219.

Bulut, H. C., Bulut, O., & Arikan, S. (2023). Evaluating group differences in online reading comprehension: The impact of item properties. International Journal of Testing, 23(1), 10–33. https://doi.org/10.1080/15305058.2022.2044821

Cano, S., & Hobart, J. (2011). The problem with health measurement. Patient Preference and Adherence, 279. https://doi.org/10.2147/PPA.S14399

Cano, S. J., Pendrill, L. R., Melin, J., & Fisher, W. P. (2019). Towards consensus measurement standards for patient-centered outcomes. Measurement, 141, 62–69. https://doi.org/10.1016/j.measurement.2019.03.056

Carlstedt, B., & Gustafsson, J.-E. (2005). Construct validation of the Swedish Scholastic Aptitude Test by means of the Swedish Enlistment Battery. Scandinavian Journal of Psychology, 13.

Carroll, J. B. (1993). Human cognitive abilities: A survey of factor-analytic studies. Cambridge University Press. https://doi.org/10.1017/CBO9780511571312

Chalmers, R. P. (2012). mirt: A multidimensional item response theory package for the R environment. Journal of Statistical Software, 48(6), 1–29. https://doi.org/10.18637/jss.v048.i06

Chan, K.-Y., Drasgow, F., & Sawin, L. L. (1999). What is the shelf life of a test? The effect of time on the psychometrics of a cognitive ability test battery. Journal of Applied Psychology, 84(4), 610–619. https://doi.org/10.1037/0021-9010.84.4.610

Christensen, K. B., Makransky, G., & Horton, M. (2017). Critical values for Yen’s Q3: Identification of local dependence in the Rasch model using residual correlations. Applied Psychological Measurement, 41(3), 178–194. https://doi.org/10.1177/0146621616677520

Clark, C., & Foster, A. (2005). Children’s and young people’s reading habits and preferences: The who, what, why, where and when. In National Literacy Trust. National Literacy Trust. https://eric.ed.gov/?id=ED541603

Dutton, E., & Lynn, R. (2013). A negative Flynn effect in Finland, 1997–2009. Intelligence, 41(6), 817–820. https://doi.org/10.1016/j.intell.2013.05.008

Embretson, S. E. (2000). Polytomous IRT. In S. E. Embretson & S. P. Reise (Eds.), Item Response Theory for psychologists (2nd ed.).

Fisher, W. P. (2009). Invariance and traceability for measures of human, social, and natural capital: Theory and application. Measurement, 42(9), 1278–1287. https://doi.org/10.1016/j.measurement.2009.03.014

Flanagan, D. P., Alfonso, V. C., Ortiz, S. O., & Dynda, A. M. (2013). Cognitive assessment: Progress in psychometric theories of intelligence, the structure of cognitive ability tests, and interpretive approaches to cognitive test performance. In The Oxford handbook of child psychological assessment (pp. 239–285). Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199796304.001.0001

Flynn, J. R. (2012). Are we getting smarter?: Rising IQ in the twenty-first century. Cambridge University Press.

Goldammer, P., Annen, H., Lienhard, C., & Jonas, K. (2024). An examination of model fit and measurement invariance of general mental ability and personality measures used in the multilingual context of the Swiss armed forces: A Bayesian structural equation modeling approach. Military Psychology, 36(1), 96–113. https://doi.org/10.1080/08995605.2021.1963632

Gustafsson, J.-E., & Balke, G. (1993). General and specific abilities as predictors of school achievement. Multivariate Behavioral Research, 28(4), 407–434. https://doi.org/10.1207/s15327906mbr2804_2

Hagquist, C., & Andrich, D. (2017). Recent advances in analysis of differential item functioning in health research using the Rasch model. Health Qual Life Outcomes 15, 181. https://doi.org/10.1186/s12955-017-0755-0

Helland-Riise, F., Norrøne, T. N., & Andersson, B. (2024). Large-scale item-level analysis of the figural matrices test in the Norwegian armed forces: Examining measurement precision and sex bias. Journal of Intelligence, 12(9). https://doi.org/10.3390/jintelligence12090082

Hobart, J., & Cano, S. (2009). Improving the evaluation of therapeutic interventions in multiple sclerosis: The role of new psychometric methods. Health Technology Assessessments. 13(12):iii, ix-x, 1-177. doi: 10.3310/hta13120.

Husén, T. (1948). Konstruktion och standardisering av svenska krigsmaktens inskrivningsprov: 1948 års version: redogörelse utarbetad vid CVB. Centrala värnpliktsbyrån.

Johansson, M. (2024a). easyRasch (Version 0.3.1) [R package]. https://github.com/pgmj/easyRasch.

Johansson, M. (2024b). Simulation based cutoff values for Rasch item fit and residual correlations. R, Rasch, Etc. https://pgmj.github.io/simcutoffs.html

Johansson, M. (2025). Detecting item misfit in Rasch models. Educational Methods & Psychometrics, 3. https://dx.doi.org/10.61186/emp.2025.5

Joint Committee on the Standards for Educational and Psychological Testing. (2014). Standards for educational and psychological testing. American Educational Research Association.

Jonsson, E., Salo, M., Lillemäe, E., Steder, F. B., Ferst, T., Kasearu, K., Novagrockiene, J., Österberg, J., Sederholm, T., Svensén, S., Tresch, T. S., & Truusa, T.-T. (2024). Multifaceted conscription: A comparative study of six European countries (1). 7(1), Article 1. https://doi.org/10.31374/sjms.166

Leitgöb, H., Seddig, D., Asparouhov, T., Behr, D., Davidov, E., De Roover, K., Jak, S., Meitinger, K., Menold, N., Muthén, B., Rudnev, M., Schmidt, P., & van de Schoot, R. (2022). Measurement invariance in the social sciences: Historical development, methodological challenges, state of the art, and future perspectives. Social Science Research, 102805. https://doi.org/10.1016/j.ssresearch.2022.102805

Lepper, C., Stang-Rabrig, J., & McElvany, N. (2022). Gender differences in reading: Examining text-based interest in relation to text characteristics and reading comprehension. Learning and Instruction, 82, 101680. https://doi.org/10.1016/j.learninstruc.2022.101680

Ludvigsson, J. F., Berglind, D., Sundquist, K., Sundström, J., Tynelius, P., & Neovius, M. (2022). The Swedish military conscription register: Opportunities for its use in medical research. European Journal of Epidemiology, 37(7), 767–777. https://doi.org/10.1007/s10654-022-00887-0

McGeown, S. P., Osborne, C., Warhurst, A., Norgate, R., & Duncan, L. G. (2016). Understanding children’s reading activities: Reading motivation, skill and child characteristics as predictors. Journal of Research in Reading, 39(1), 109–125. https://doi.org/10.1111/1467-9817.12060

McGrew, K. S. (2009). CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research. Intelligence, 37(1), 1–10. https://doi.org/10.1016/j.intell.2008.08.004

McNamara, T., & Knoch, U. (2012). The Rasch wars: The emergence of Rasch measurement in language testing. Language Testing, 29(4), 555–576. https://doi.org/10.1177/0265532211430367

Meisenberg, G., & Lynn, R. (2023). Ongoing trends of human intelligence. Intelligence, 96, 101708. https://doi.org/10.1016/j.intell.2022.101708

Melin, J., Cano, S., Flöel, A., Göschel, L., & Pendrill, L. (2022). The role of entropy in construct specification equations (CSE) to improve the validity of memory tests: Extension to word lists. Entropy, 24(7). https://doi.org/10.3390/e24070934

Messick, S. (1995). Validity of psychological assessment: Validation of inferences from persons’ responses and performances as scientific inquiry into score meaning. American Psychologist, 50(9), 741–749. https://doi.org/10.1037/0003-066X.50.9.741

Morel, T., & Cano, S. J. (2017). Measuring what matters to rare disease patients – reflections on the work by the IRDiRC taskforce on patient-centered outcome measures. Orphanet Journal of Rare Diseases, 12, 171. https://doi.org/10.1186/s13023-017-0718-x

Muthukrishna, M., & Henrich, J. (2016). Innovation in the collective brain. Philosophical Transactions of the Royal Society B: Biological Sciences, 371(1690), 20150192. https://doi.org/10.1098/rstb.2015.0192

Nordlund, P. (2022). Sweden and Swedish defence – Introduction to the special issue. Defence and Peace Economics, 33(4), 387–398. https://doi.org/10.1080/10242694.2021.2003529

Nordmo, M., Norrøne, T. N., & Lang-Ree, O. C. (2025). Reevaluating the Flynn effect, and the reversal: Temporal trends and measurement invariance in Norwegian armed forces intelligence scores. Intelligence, 110, 101909. https://doi.org/10.1016/j.intell.2025.101909

Pendrill, L. (2014). Man as a measurement instrument. NCSLI Measure, 9(4), 24–35. https://doi.org/10.1080/19315775.2014.11721702

Rasch, G. (1960). Studies in mathematical psychology: I. Probabilistic models for some intelligence and attainment tests. Nielsen & Lydiche.

Reckase, M. D. (2009). Unidimensional item response theory models. In M. D. Reckase (Ed.), Multidimensional item response theory (pp. 11–55). Springer. https://doi.org/10.1007/978-0-387-89976-3_2

Roebuck-Spencer, T. M., Reeves, D. L., Bleiberg, J., Cernich, A. N., Schwab, K., Ivins, B., Salazar, A., Harvey, S., Brown, F., & Warden, D. (2008). Influence of demographics on computerized cognitive testing in a military sample. Military Psychology, 20(3), 187–203. https://doi.org/10.1080/08995600802118825

Rönnlund, M., Carlstedt, B., Blomstedt, Y., Nilsson, L.-G., & Weinehall, L. (2013). Secular trends in cognitive test performance: Swedish conscript data 1970–1993. Intelligence, 41(1), 19–24. https://doi.org/10.1016/j.intell.2012.10.001

Shiu, W., Beaujean, A. A., Must, O., te Nijenhuis, J., & Must, A. (2013). An item-level examination of the Flynn effect on the national intelligence test in Estonia. Intelligence, 41(6), 770–779. https://doi.org/10.1016/j.intell.2013.05.007

Sijtsma, K., & Junker, B. W. (2006). Item response theory: Past performance, present developments, and future expectations. Behaviormetrika, 33, 75–102. https://doi.org/10.2333/bhmk.33.75

Skolverket. (2023). PIRLS 2021 Läsförmågan hos svenska elever i årskurs 4 i ett internationellt perspektiv (SKOLV-R-2023:4-SE). Skolverket. https://www.skolverket.se/publikationer?id=11490

Smith, R. (1993). Guessing and the Rasch model. Rasch Measurement Transactions, 6(4), 262–263.

Somaraju, A. V., Nye, C. D., & Olenick, J. (2022). A review of measurement equivalence in organizational research: what’s old, what’s new, what’s next? Organizational Research Methods, 25(4), 741–785. https://doi.org/10.1177/10944281211056524

Stark, S., Chernyshenko, O. S., & Drasgow, F. (2004). Examining the effects of differential item (functioning and differential) test functioning on selection decisions: When are statistically significant effects practically important? Journal of Applied Psychology, 89(3), 497–508. https://doi.org/10.1037/0021-9010.89.3.497

Stenner, A. J., Burdick, H., Sanford, E. E., & Burdick, D. S. (2006). How accurate are lexile text measures? Journal of Applied Measurement, 7(3), 307–322.

Strobl, C., Kopf, J., & Zeileis, A. (2015). Rasch trees: A new method for detecting Differential Item Functioning in the Rasch model. Psychometrika, 80(2), 289–316. https://doi.org/10.1007/s11336-013-9388-3

Sveriges kommuner och landsting. (2019). Könsskillnader i skolresultat. Sveriges kommuner och landsting. https://skr.se/skr/tjanster/rapporterochskrifter/publikationer/konsskillnaderiskolresultat.65164.html

Svetina, D., Gorin, J. S., & Tatsuoka, K. K. (2011). Defining and comparing the reading comprehension construct: A cognitive-psychometric modeling approach. International Journal of Testing, 11(1), 1–23. https://doi.org/10.1080/15305058.2010.518261

Ullstadius, E., Carlstedt, B., & Gustafsson, J.-E. (2008). The multidimensionality of verbal analogy items. International Journal of Testing, 8(2), 166–179. https://doi.org/10.1080/15305050802001243

Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3(1), 4–70. https://doi.org/10.1177/109442810031002

Williams, R. L. (2013). Overview of the Flynn effect. Intelligence, 41(6), 753–764. https://doi.org/10.1016/j.intell.2013.04.010

Wilson, M. (2013). Seeking a balance between the statistical and scientific elements in psychometrics. Psychometrika, 78(2), 211–236. https://doi.org/10.1007/s11336-013-9327-3