LIWC and the TAT.

Abstract

A Fielding Graduate University paper makes the case that Linguistic Inquiry and Word Count belongs in Thematic Apperception Test interpretation, and an accompanying pipeline scores 677 card responses from 69 participants across 31 TAT cards to test whether specific cards reliably pull for the themes they were designed to elicit; together they sketch a path toward continuous, item-level TAT scoring.

Function words as a second channel.

The Thematic Apperception Test asks a respondent to look at an ambiguous picture and tell a story about it1Morgan & Murray (1935). A method for investigating fantasies: The Thematic Apperception Test. Archives of Neurology & Psychiatry, 34, 289-306. The projective hypothesis holds that ambiguous stimuli are perceived and organized according to the respondent’s needs, motives, and feelings (Teglasi, 2010)., and traditional interpretation has tended to treat those stories as content, with a clinician reading the narrative for themes, conflicts, object relations, and the rest; the recurring weakness of the method, ever since Murray first published it in 1935, has been inter-rater variability.

Tausczik and Pennebaker (2010) draw a distinction that seems worth carrying into projective assessment. Content words carry the subject matter, while function words2Tausczik & Pennebaker (2010). The psychological meaning of words: LIWC and computerized text analysis methods. Journal of Language and Social Psychology, 29(1), 24-54. Function words are pronouns, prepositions, articles, conjunctions, and auxiliary verbs, the structural scaffolding of a sentence. carry the structural scaffolding, governing how sentences connect and how the speaker positions themselves relative to others, and these tend to sit largely outside conscious management; a respondent can suppress mention of family or money, and yet they cannot so easily suppress how often they reach for the word “I.”

LIWC2015 is a deterministic dictionary parser that scores a text against roughly ninety psychometrically anchored categories, pronoun families, affect categories, social processes, cognitive processes, drives, and personal concerns among them. Pairing it with the TAT does not stand in for the clinician; I built the pipeline to add a second channel of evidence, one that answers to the parts of language the respondent is not watching.

Six hundred seventy-seven narratives, split four ways.

The corpus is built from a TAT administration to 69 participants covering 31 card types; each narrative was cleaned to strip assessor prompts and behavioral observations, then split so the same dataset can be queried at four different grains, namely each individual response to each individual card, each individual’s full protocol, all responses to a single card, and the whole corpus read as a single document. LIWC-published base rates for blogs, expressive writing, novels, and other text genres sit alongside the TAT data, so any finding can be sanity-checked against an external reference.

Card responses

677

Independent card-level narratives from 69 participants, after cleaning and exclusion of mislabeled records.

Cards represented

Frequencies range from 60 (Card 2) down to 2 (Card 13G); card-level analysis is limited to the higher-frequency cards.

LIWC2015 features

~90

Pronoun families, affect categories, social processes, cognitive processes, drives, personal concerns, temporal focus.

Card-level analysis leans toward the cards that produced enough narratives to support inference, and the dataset assigns five mutually exclusive types to each row (INDVNARR, INDVCARDALLRESP, INDVRESPALLCARDS, WHOLECORPORA, BASERATE), so the same LIWC variable can be examined at whichever grain the question happens to require.

Cards pull for the themes they were designed to elicit.

A MANOVA across Cards 1, 2, 3BM, 4, and 8BM (the five highest-frequency cards) asks whether LIWC variables differ by card, and they do, with significance on every emotion, social, drive, and personal concern category examined; the signature each card produces lines up with what the test’s clinical literature claims it elicits.

Card 1 (boy with violin) pulls 1.88x the corpus baseline for achievement language, Card 2 (farm scene) pulls 1.80x for family and 2.55x for work, Card 3BM (huddled figure) pulls 4.74x for sadness and 2.17x for anxiety, Card 4 (woman restraining a man) pulls 3.38x for anger, and Card 8BM (surgical scene) pulls 4.57x for death language. Every contrast across cards reaches p < .05, and most reach p < .001.

LIWC Means by Card, normalized to per-axis max

Card 1 / Achievement Card 3BM / Sadness Card 8BM / Death

Radar chart showing six LIWC axes, namely anxiety, anger, sadness, achievement drive, power drive, and death, for each of the five TAT cards analyzed (Cards 1, 2, 3BM, 4, and 8BM), with values normalized to the per-axis maximum across cards.

Six LIWC axes per card, namely anxiety, anger, sadness, achievement drive, power drive, and death. Values are mean LIWC scores from the MANOVA, normalized to the per-axis maximum across the five compared cards (Cards 1, 2, 3BM, 4, and 8BM).

Pronouns and negation as clinical signal.

Function-word patterns map onto clinical states with a consistency that tends to surprise; depressed and suicidal individuals use more first-person singular pronouns, more death-related words, and more negatively valenced language (Bernard et al., 2016), grief over a romantic loss raises first-person singular and lowers causal words (Boals & Klein, 2005), and high rates of negation track with the self-aggression associated with suicidality and substance dependence (Hargitai et al., 2007). The reading has to stay careful, since a high first-person singular can mark pain, or pathological egocentrism, or simple self-awareness, depending on what else is going on in the language.

Psychotic speech leaves a different signature, and the open-narrative format of the TAT lets neologisms, tangentiality, and the kind of temporal-proximity word linkage central to the Cloze procedure (Salzinger et al., 1964) surface in a way structured tests cannot quite capture; pairing LIWC with TAT narratives helps separate disordered thought form from disordered thought content, a distinction that drives both conceptualization and intervention (Langdon, Coltheart, Ward, & Catts, 2002).

Trait measurement gains here as well; linguistic analysis of narratives can tell person-oriented from thing-oriented respondents (McIntyre & Graziano, 2016), and latent semantic analysis recovers Big Five traits from natural speech (Kwantes et al., 2016), while for cognitive functioning the linguistic content of TAT narratives discriminates participants with agenesis of the corpus callosum from age- and intelligence-matched controls (Turk, Brown, Symington, & Paul, 2010), which points toward a use case beyond the affective domain.

Toward continuous item-level TAT scoring.

The card-level effects in section 03 suggest a way to put a little more structure on TAT interpretation; if each card pulls reliably for a known set of LIWC variables, then those variables can serve as the card’s measurement axes, and a respondent’s score on each axis can be weighed against a card-specific norm metric rather than a global one. Is this person’s response to Card 3BM unusually sad compared to other people’s responses to Card 3BM? Does the same elevation show up across their full protocol? Does it line up with a current or historical diagnosis?

The framing sits closer to continuous manifest-variable item response than to traditional projective scoring, with each card treated as its own measure carrying its own threshold and the whole protocol treated as a larger measure built up from those item-level scores. The clinician stays in the loop throughout, since LIWC cannot read irony, idiom, or the contextual oddness of a narrative that draws clinical attention through its strangeness rather than its frequency counts; what the function-word channel adds is a second set of numbers the clinician can check a reading against. The respondent cannot watch the words they are not watching, and that is exactly where the second channel earns its keep.

References

Aronow, E., Weiss, K. A., & Reznikoff, M. (2001). A practical guide to the Thematic Apperception Test: The TAT in clinical practice. Routledge.
Bernard, J. D., Baddeley, J. L., Rodriguez, B. F., & Burke, P. A. (2016). Depression, language, and affect: An examination of the influence of baseline depression and affect induction on language. Journal of Language and Social Psychology, 35(3), 317-326.
Boals, A., & Klein, K. (2005). Word use in emotional narratives about failed romantic relationships and subsequent mental health. Journal of Language and Social Psychology, 24(3), 252-268.
Cohen, A. S., Minor, K. S., Baillie, L. E., & Dahir, A. M. (2007). Clarifying the linguistic signature: Measuring personality from natural speech. Journal of Personality Assessment, 90(6), 559-563.
Gruber, N., & Kreuzpointner, L. (2013). Measuring the reliability of picture story exercises like the TAT. PLoS One, 8(11), e79450.
Hargitai, R., Naszódi, M., Kis, B., Nagy, L., Bóna, A., & László, J. (2007). Linguistic markers of depressive dynamics in self-narratives: Negation and self-reference. Empirical Text and Culture Research, 3, 26-38.
He, Q., Veldkamp, B. P., Glas, C. A., & de Vries, T. (2015). Automated assessment of patients’ self-narratives for posttraumatic stress disorder screening using natural language processing and text mining. Assessment, 1-16.
Kuperberg, G. R. (2010). Language in schizophrenia part 1: An introduction. Language and Linguistics Compass, 4(8), 576-589.
Kwantes, P. J., Derbentseva, N., Lam, Q., Vartanian, O., & Marmurek, H. H. (2016). Assessing the Big Five personality traits with latent semantic analysis. Personality and Individual Differences, 102, 229-233.
Langdon, R., Coltheart, M., Ward, P. B., & Catts, S. V. (2002). Disturbed communication in schizophrenia: The role of poor pragmatics and poor mind-reading. Psychological Medicine, 32, 1273-1284.
Leffew, H. (2017). Interpretation of the Thematic Apperception Test: A case for the inclusion of linguistic analysis. Fielding Graduate University.
Marini, A., Spoletini, I., Rubino, I. A., Ciuffa, M., Bria, P., Martinotti, G., … Spalletta, G. (2008). The language of schizophrenia: An analysis of micro and macrolinguistic abilities and their neuropsychological correlates. Schizophrenia Research, 105(1-3), 144-155.
McIntyre, M. M., & Graziano, W. G. (2016). Seeing people, seeing things: Individual differences in selective attention. Personality and Social Psychology Bulletin, 42(9), 1258-1271.
Morgan, C. D., & Murray, H. A. (1935). A method for investigating fantasies: The Thematic Apperception Test. Archives of Neurology & Psychiatry, 34, 289-306.
Pennebaker, J. W., Booth, R. J., Boyd, R. L., & Francis, M. E. (2015). Linguistic Inquiry and Word Count: LIWC2015. Pennebaker Conglomerates.
Pennebaker, J. W., & King, L. A. (1999). Linguistic styles: Language use as an individual difference. Journal of Personality and Social Psychology, 77(6), 1296-1312.
Salzinger, K., Portnoy, S., & Feldman, R. S. (1964). Verbal behavior of schizophrenic and normal subjects. Annals of the New York Academy of Sciences, 105, 845-860.
Seih, Y. T., Beier, S., & Pennebaker, J. W. (2016). Development and examination of the linguistic category model in a computerized text analysis method. Journal of Language and Social Psychology, 1-13.
Semin, G. R., & Fiedler, K. (1988). The cognitive functions of linguistic categories in describing persons: Social cognition and language. Journal of Personality and Social Psychology, 54, 558-568.
Tausczik, Y. R., & Pennebaker, J. W. (2010). The psychological meaning of words: LIWC and computerized text analysis methods. Journal of Language and Social Psychology, 29(1), 24-54.
Teglasi, H. (2010). Essentials of TAT and other storytelling assessments (2nd ed.). John Wiley & Sons.
Turk, A. A., Brown, W. S., Symington, M., & Paul, L. K. (2010). Social narratives in agenesis of the corpus callosum: Linguistic analysis of the Thematic Apperception Test. Neuropsychologia, 48(1), 43-50.
Winter, D. G. (1998). “Toward a science of personality psychology”: David McClelland’s development of empirically derived TAT measures. History of Psychology, 1(2), 130-153.