Home

This site hosts Automated Integrative Complexity, a computerized framework to rapidly code Integrative Complexity.

The system was developed by Professor Lucian Gideon Conway III (PhD) and Kathrene R. Conway.

The only qualifications for academic users are: that you have an institutional affiliation; and, should you publish something that includes data generated by the Automated Integrative Complexity system, that you cite the following three papers which formally introduced AutoIC to the academic research community:

  • Conway, L. G., III, Conway, K. R., Gornick, L. J., & Houck, S. C. (2014). Automated integrative complexity. Political Psychology, 35, 603-624. DOI:10.1111/pops.12021

To begin, select one of the following user types:

Past Research With AutoIC

Abe, J. A. A. (2022). Cognitive-Affective Styles of Biden and Trump Supporters: An Automated Text Analysis StudySocial Psychological and Personality Science, 19485506221082737.

Karvetski, C. W., Meinel, C., Maxwell, D. T., Lu, Y., Mellers, B. A., & Tetlock, P. E. (2022). What do forecasting rationales reveal about thinking patterns of top geopolitical forecasters?International Journal of Forecasting38(2), 688-704.

Conway, L. G., III, & Zubrod, A. (2022). Are American Presidents becoming less rhetorically complex? Evaluating the integrative complexity of Joe Biden and Donald Trump in historical context. Journal of Language and Social Psychology. 

Aleksovska, M. (2021). Accountable for What? The Effect of Accountability Standard Specification on Decision-Making Behavior in the Public SectorPublic Performance & Management Review44(4), 707-734.

Aleksovska, M., & Schillemans, T. (2021). Dissecting multiple accountabilities: A problem of multiple forums or of conflicting demands?Public Administration.

Arrowood, R. B., Vail III, K. E., & Cox, C. R. (2021). The existential quest: Doubt, openness, and the exploration of religious uncertaintyThe International Journal for the Psychology of Religion, 1-38.

Conway III, L. G., McFarland, J. D., Costello, T. H., & Lilienfeld, S. O. (2021). The curious case of left‐wing authoritarianism: When authoritarian persons meet anti‐authoritarian normsJournal of Theoretical Social Psychology.

Gallacher, J. D., Heerdink, M. W., & Hewstone, M. (2021). Online engagement between opposing political protest groups via social media is linked to physical violence of offline encountersSocial Media+ Society7(1), 2056305120984445.

Gregory, A. L., & Piff, P. K. (2021). Finding uncommon ground: Extremist online forum engagement predicts integrative complexityPlos one16(1), e0245651.

Kara-Yakoubian, M., Meyers, E. A., Sharpinskyi, K., Dorfman, A., & Grossmann, I. (2021). Hidden wisdom or pseudo-profound bullshit? The effect of speaker admirability. https://doi.org/10.31234/osf.io/tpnkw

McCullough, H. (2021). Green slime and orange blimps: A linguistic analysis of Nickelodeon’s kids’ choice awardsPsychology of Aesthetics, Creativity, and the Arts. Advance online publication.

McCullough, H. (2021). Integrative Complexity, Horror, and GenderPress Start7(1), 1-18.

McCullough, H. (2021). Be complex, be very complex: Evaluating the integrative complexity of main characters in horror filmsPsychology of Popular Media, 10(1), 50–58.

Mell, J. N., Jang, S., & Chai, S. (2021). Bridging temporal divides: Temporal brokerage in global teams and its impact on individual performanceOrganization Science32(3), 731-751.

Nosrat, A. (2021). Middle manager cognition and socio-environmental performance: the case of climate change. Doctoral dissertation, McGill University.

Suedfeld, P., Morrison, B. H., & Kuznar, L. A. (2021). National interests and the Trump Doctrine: the meaning of “America First”. In The Trump Doctrine and the Emerging International System (pp. 39-70). Palgrave Macmillan, Cham.

Woodard, S. R., Chan, L., & Conway III, L. G. (2021). In search of the cognitively complex person: Is there a meaningful trait component of cognitive complexity? Personality and Social Psychology Review25(2), 95-129.

Zubrod, A., Conway III, L. G., Conway, K. R., & Ailanjian, D. (2021). Understanding the role of linguistic complexity in famous trial outcomesJournal of Language and Social Psychology40(3), 354-377.

Conway, L. G., III, Conway, K. R., & Houck, S. C. (2020). Validating Automated Integrative Complexity: Natural language processing and the Donald Trump test. Journal of Social and Political Psychology, 8, 504-524.

Conway III, L. G., & Woodard, S. R. (2020). Integrative complexity across domains and across time: Evidence from political and health domainsPersonality and Individual Differences155, 109713.

McCullough, H. (2020). The diamonds and the dross: A quantitative exploration of integrative complexity in fanfictionPsychology of Popular Media9(1), 59-68.

Sharp, C. A., Shariff, A. F., & LaBouff, J. P. (2020). Religious complexity and intergroup biasThe International Journal for the Psychology of Religion30(2), 73-88.

Gallacher, J. D., & Heerdink, M. W. (2019). Measuring the effect of Russian Internet research agency information operations in online conversationsDefence Strategic Communications6, 155-198.

McCullough, H. (2019). Hey! Listen!Press Start5(1), 94-107.

McCullough, H. (2019). From Zelda to StanleyPress Start5(2), 137-149.

McCullough, H., & Kalsher, M. J. (2019). The integrative complexity of media reports of natural disasters: A preliminary analysis. In Proceedings of the Human Factors and Ergonomics Society Annual Meeting (Vol. 63, No. 1, pp. 601-605). Sage CA: Los Angeles, CA: SAGE Publications.

Weeks, M., & Geisler, S. (2019). Revisiting cognitive complexity of religious topics: Multiple complexity, religious fundamentalism, and a quest orientationPsychology of Religion and Spirituality, 11(4), 433–441.

Houck, S. C., Conway, L. G. III, Parrow, K., & Luce, A., & Salvati, J. (2018). An integrative complexity analysis of religious and irreligious thinkingSage Open. https://doi.org/10.1177/2158244018796302

Salvati, J. M., & Houck, S. C. (2019). Examining the causes and consequences of confession-eliciting tactics during interrogationJournal of Applied Security Research14(3), 241-256.

Williams, B. (2018). Group heterogeneity influences complex and novel outcomes. Doctoral dissertation, The University of Texas at Arlington.

McCullough, H. (2018). The new American dream: YouTube, movie review and integrative complexity. In 6th Annual University of Albany—SUNY Institutions & Societies Conference, Albany, NY.

McCullough, H., & Conway, L. G., III.  (2018). The cognitive complexity of Miss Piggy and Osama Bin Laden: Examining linguistic differences between fiction and realityPsychology of Popular Media Culture, 7, 518-532.

McCullough, H., & Conway, L. G., III.  (2018). “And the Oscar goes to…”: Integrative complexity’s predictive power in the film industry. Psychology of Aesthetics, Creativity, and the Arts, 12, 392-398.

Felts, N. A. (2017). Please explain yourself: Mechanisms of opinion improvement in deliberative forums. Doctoral dissertation, The Ohio State University.

Prinsloo, C. F. (2017). Investigating the influence of individual value systems and risk propensities on decision-making quality in value clashing circumstances. Doctoral dissertation, University of Pretoria.

Putra, I. E., Erikha, F., Arimbi, R. S., & Rufaedah, A. (2018). Increasing integrative complexity on convicted terrorists in IndonesiaSocial Psychology and Society9(2), 35-45.

Houck, S. C., Repke, M. A., & Conway, L. G., III.  (2017). Understanding what makes terrorist groups’ propaganda effective: An integrative complexity analysis of ISIL and Al QaedaJournal of Policing, Intelligence and Counter Terrorism, 12, 105-118. DOI:10.1080/18335330.2017.1351032

Kraichy, D. (2016). Making sense of high potential, talent, and leadership in organizations: a discursive and psychological approach. Master’s thesis, University of Manitoba.

Skillicorn, D. B. (2015). Empirical assessment of Al Qaeda, ISIL, and Taliban propaganda. In 2015 IEEE international conference on intelligence and security informatics (ISI) (pp. 61-66). IEEE.

Skillicorn, D. B., & Reid, E. F. (2014). Language use in the Jihadist magazines inspire and AzanSecurity Informatics3(1), 1-16.

Conway, L. G., III, Conway, K. R., Gornick, L. J., & Houck, S. C. (2014). Automated integrative complexity. Political Psychology, 35, 603-624.

 

Upload

Before running analysis, please make sure that you’ve read the instructions. If you are using the AutoIC Paragraph version, make sure you download the CSV template below.

On uploading material, you also confirm that the necessary legal permissions are in place for material sent through the system.

If you are planning on running a larger project (greater than 30 files at once), please contact us.  We need to allot time for processing large batches of files.   We can also setup an easier method for uploading and distributing files which will save you from having to upload each file individually.

AutoIC Document Analysis

Take prepared material and upload it below:

    • Single TXT file you would like to run analysis on, or
    • ZIP file containing any number of TXT files you would like to code (However, zip files should be less than 1MB to process well.  One .5MB file, will require about 4 hours to run.)

Note to Mac users: When you zip your file, a __MACOSX directory is automatically created in the zip file. Please delete this folder before uploading.

AutoIC Paragraph Analysis

  1. Download the CSV Template
  2. The CSV has a single sheet. There are five column headers. The first three columns (titled respectively, “Var1,” Var2,” and “Var3.”) are columns for variable names which you may or may not use as you see fit. The fourth column, titled “Paragraph,” is where you need to paste the things you want code.  Please note:  The file will stop coding when it reaches the first row with no data in the first four columns. However, if there is data in columns 1-3, but nothing in column 4, your file will not run.
  3. Take your prepared version of the CSV template file and upload it below

Please select your file:

Upload files

Press the ‘Upload File’ button above.

You can upload as many files as you need run, and we will process them as soon as we can.  When your analysis is complete, you will receive an email(s) with a spreadsheet of your results.  If you have uploaded files, and haven’t received the coded response within a week, feel free to contact me at kathrene.conway@umontana.edu.


Not sure what type of analysis to use or what the variables are? Read-up on the difference on the instructions page.

By uploading any files and using the system, you are agreeing to the terms of use.

Instructions

Below is a brief overview of Automated Integrative Complexity.


Documents or Paragraphs?
There are two principle ways to run material through AutoIC:

  1. By Document
  2. By Paragraph

AutoIC for Documents

  • AutoIC for documents takes a complete user provided document and codes the entire thing. It does this by breaking the document into optimal 75-word chunks of text. If you have multiple documents, or large pieces of text, you should use the Document version.  Please note that the system, because of its multi-pass process, will slow down tremendously with large zipped batches.  A .5MB zipped document will take a couple of hours to process.  We recommend zipping larger projects into batches of about this size.

AutoIC for Paragraphs

  • AutoIC for Paragraphs scores whole paragraphs without breaking them up. The primary purpose we had in mind for the Paragraph version was an easy way to score participant responses in laboratory settings. You should only use the paragraph format for relatively short segments. This system is designed to score things in a paragraph-by-paragraph format.  Please note:  The file will stop coding when it reaches the first blank field in Column 4.

Why Documents or Paragraphs?
If you have a document that is, say, 2000 words long, and you use the Paragraph analysis, it will likely assign it a very high score; the Paragraph version was designed for shorter, paragraph-length responses. There is a well know ‘length effect’ when coding Integrative Complexity. Using the Document version of the system avoids this problem. Of course, you are free to use either as long as you meet the terms of use. The only thing that is different is the unit of scoring – that is, how it decides which words to score. See the Automated Integrative Complexity Manual for more details.

Variables on the PARAGRAPH SHEET:

  • Document = name of document from which chunk was scored
  • Chunk = chunk number within document; always listed sequentially (so 1 = first chunk in document, 2 = second, and so forth)
  • Complete? = did the chunk include a full 75 words (=1), or not (=0).  Every chunk will include 75 words except for the last chunk in a document.  AutoIC scores the last chunk, but this variable allows researchers to select out those “incomplete” chunks easily and do analyses without them should they choose to do so.  There are pros and cons to both approaches: (1) If you don’t remove them, you have chunks in your data that are not standard length; (2) but if you do remove them, you systematically lose the end of all documents. (In most datasets, it will make little or no difference; more on this when we discuss the document-level variables).  On balance, though, our advice is to remove the incomplete chunks.  On average, they are just adding unnecessary noise in terms of paragraphs with differing word lengths.  And you’re still going to be scoring the vast majority of the words in most cases anyway.  If it looks like you are not (see DOCUMENT SHEET below), then you should just do it both ways (look at your data with and without incomplete chunks).  If you do, you’ll almost certainly see what I’m telling you here: It just doesn’t matter.
  • Paragraph = the actual paragraph that was scored, devoid of punctuation.
  • Words = number of words in the paragraph; it’s always 75 except for the last chunk in each document.
  • IC = integrative complexity score for that chunk.
  • DIAL = dialectical complexity score for that chunk.
  • ELAB = elaborative complexity score for that chunk.
  • IC_Differentiation = Integrative complexity score that was the result of differentiation.  Note that, to keep differentiation and integration on the same scale, neither of them starts from 1 (they both start from zero).  The IC scale itself starts from 1.  I point this out only so that you will realize that IC_Differentiation + IC_Integration does NOT = IC.  Rather, you have to add one to that score to get the IC score.  This is just to keep the differentiation and integration scores on the same scale.
  • IC_Integration = Integrative complexity score that was the result of integration. See note for IC_Differentiation about the scale.  This is also true for all the differentiation and integration variables below; we’re just going to stop saying that.
  • DIAL_Differentiation = Dialectical complexity score that was the result of differentiation.
  • DIAL_Integration = Dialectical complexity score that was the result of integration. See note for IC-Integration above.
  • ELAB_Differentiation = Elaborative complexity score that was the result of differentiation.
  • ELAB_Integration = Elaborative complexity score that was the result of integration. See note for IC-Integration above.
  • ELAB = elaborative complexity score for that chunk.
  • IC_Differentiation = Integrative complexity score that was the result of differentiation.  Note that, to keep differentiation and integration on the same scale, neither of them starts from 1 (they both start from zero).  The IC scale itself starts from 1.  I point this out only so that you will realize that IC_Differentiation + IC_Integration does NOT = IC.  Rather, you have to add one to that score to get the IC score.  This is just to keep the differentiation and integration scores on the same scale.
  • IC_Integration = Integrative complexity score that was the result of integration. See note for IC_Differentiation about the scale.  This is also true for all the differentiation and integration variables below; we’re just going to stop saying that.
  • DIAL_Differentiation = Dialectical complexity score that was the result of differentiation.
  • DIAL_Integration = Dialectical complexity score that was the result of integration. See note for IC-Integration above.
  • ELAB_Differentiation = Elaborative complexity score that was the result of differentiation.
  • ELAB_Integration = Elaborative complexity score that was the result of integration. See note for IC-Integration above.

Variables on the DOCUMENT SHEET:

  • Document = name of document.
  • Num Complete Chunks = number of complete chunks in the document.
  • Num Words in Complete Chunks = Total number of words in the document, minus the incomplete chunks.
  • Percentage of Words Scored = This variable is only relevant to the main set of IC variables (which remove incomplete chunks).  For those main IC variables, this tells you, out of all the words in the document, what percentage was actually scored (this will be all of the words except for incomplete chunks).  Unless you have really short documents, this figure is going to be above 90%, and usually will hover above 95%.

The next set of variables parallels those for the paragraph sheet, only it removes incomplete chunks.  This is the set we recommend using for the typical data set:

  • IC = average integrative complexity score for the document, excluding incomplete chunks .
  • DIAL = average dialectical complexity score for the document, excluding incomplete chunks.
  • ELAB = average elaborative complexity score for the document excluding incomplete chunks.
  • IC_Differentiation = average level of IC differentiation for that document (please see notes on integration/differentiation for paragraph sheet), excluding incomplete chunks.
  • IC_Integration = average level of IC integration for that document (please see notes on integration/differentiation for paragraph sheet), excluding incomplete chunks.
  • DIAL_Differentiation = average level of DIAL differentiation for that document (please see notes on integration/differentiation for paragraph sheet), excluding incomplete chunks.
  • DIAL_Integration = average level of DIAL integration for that document (please see notes on integration/differentiation for paragraph sheet), excluding incomplete chunks.
  • ELAB_Differentiation = average level of ELAB differentiation for that document (please see notes on integration/differentiation for paragraph sheet), excluding incomplete chunks.
  • ELAB_Integration = average level of ELAB integration for that document (please see notes on integration/differentiation for paragraph sheet), excluding incomplete chunks.

The next set of variables parallels those for the paragraph sheet, only it KEEPS incomplete chunks:

  • Num All Chunks = Total number of words in the document, including the incomplete chunks.
  • Mean Words in Chunk = average number of words in the chunk; again, if it drifts far below 70, this suggests a higher percentage of your paragraphs are below the 75 marker.
  • IC_AllChunks = average integrative complexity score for the document, including incomplete chunks.
  • DIAL_AllChunks = average dialectical complexity score for the document, including incomplete chunks.
  • ELAB_AllChunks = average elaborative complexity score for the document, including incomplete chunks.
  • DIAL_AllChunks = average dialectical complexity score for the document, including incomplete chunks.
  • ELAB_AllChunks = average elaborative complexity score for the document, including incomplete chunks.
  • IC_Differentiation_AllChunks = average level of differentiation for that document (please see notes on integration/differentiation for paragraph sheet), including incomplete chunks.
  • IC_Integration_AllChunks = average level of integration for that document (please see notes on integration/differentiation for paragraph sheet), including incomplete chunks.
  • DIAL_Differentiation_AllChunks = average level of dialectical differentiation for that document (please see notes on integration/differentiation for paragraph sheet), including incomplete chunks.
  • DIAL_Integration_AllChunks = average level of dialectical integration for that document (please see notes on integration/differentiation for paragraph sheet), including incomplete chunks.
  • ELAB_Differentiation_AllChunks = average level of elaborative differentiation for that document (please see notes on integration/differentiation for paragraph sheet), including incomplete chunks.
  • ELAB_Integration_AllChunks = average level of elaborative integration for that document (please see notes on integration/differentiation for paragraph sheet), including incomplete chunks.

 

Each user of the system confirms that they have the necessary legal permissions in place for material sent through the system.

Terms of Use

Automated Integrative Complexity (AutoIC), was originally developed by Professor Lucian Gideon Conway III, PhD, and Kathrene R. Conway. It appears here with Kathrene Conway’s and Lucian Conway’s endorsement and support. Thousands of hours and dozens of people where involved in the conception, design, build, and testing of the original system. One of the core elements of these efforts has been to make AutoIC available for unrestricted academic use.

The system is copyrighted and provided for free to the academic research community.

The user also confirms that they have the necessary legal permissions in place for material sent through the system.

Academic Users

The only qualifications for academic users are: that you have an institutional affiliation; and, should you publish something that includes data generated by the Automated Integrative Complexity system, that you cite the following three papers which formally introduced and validated AutoIC in the academic research community: