Open Access Repositorium für Messinstrumente


Partisan Identity Scale (PIS)

  • Autor/in: Bankert, A. & Huddy, L.
  • In ZIS seit: 2019
  • DOI:
  • Abstract: The partisan identity scale is based on the Identification with a Psychological Group (IDPG) scale created by Mael and Tetrick (1992). We adapted five items from the original scale, and added t ... mehrhree new items for inclusion in national surveys in the U.K., Sweden, and the Netherlands. We primarily utilize the scale to predict in-party vote and political participation. The scale targets partisans only. weniger
  • Sprache Dokumentation: English
  • Sprache Items: englisch
  • Anzahl der Items: 8
  • Erhebungsmodus: CAWI
  • Bearbeitungszeit: on average 1 minute and 20 seconds (10 seconds per item)
  • Reliabilität: Cronbach's Alpha = .83 to .86
  • Validität: evidence for criterion validity
  • Konstrukt: expressive partisanship
  • Schlagwörter: identity, partisanship, party Affiliation
  • Item(s) in Bevölkerungsumfrage eingesetzt: yes
  • URL Datenarchiv: 
  • Entwicklungsstand: validated, standardized
    • Instruction

      Introductory phrase: "How frequently do you experience the following thoughts, feelings, and behaviors?"


      Items can be placed in a grid whereby each item's position in the grid randomly varies to avoid order effects. In each item, insert the name of the party that the respondent identifies with.








      When I speak about this party, I usually say "we" instead of "they".




      I am interested in what other people think about this party.




      When people criticize this party, it feels like a personal insult.


      Included in shorter version of scale.


      I have a lot in common with other supporters of this party.




      If this party does badly in opinion polls, my day is ruined.




      When I meet someone who supports this party, I feel connected.


      Included in shorter version of scale.


      When I speak about this party, I refer to them as "my party".


      Included in shorter version of scale.


      When people praise this party, it makes me feel good.


      Included in shorter version of scale.


      Response specifications

      Each of the 8 items was presented with the four response categories 1 = "Never/Rarely", 2 = "Sometimes", 3 = "Often", and  4 = "Always".



      The scale score is computed as an unweighted mean of all items. Higher values indicate stronger partisan identity. If space is limited in the survey, a shorter subscale comprised of items 3, 6, 7, and 8 can be utilized. Non-responses can either be set to missing values or replaced with imputed values.


      Application field

      The scale measures partisan identity strength - the degree to which voters identify with a political party. Since partisan identity is a latent concept, the scale consists of several items to fully capture even fine variations in identity strength. The scale is normally distributed in online surveys but since no item appears to be overly sensitive, it can also be used in oral survey. The scale targets partisans only. It cannot be used for respondents who report no attachment to any political party. Due to its short completion time (on average 1 minute and 20 seconds; 10 seconds per item) the instrument can be applied in research settings with severe time restrictions, for example, large-scale surveys.


    The scale originates in the expressive identity approach to partisanship, which has gained credence in the U.S. as well as European multi-party systems (see Bankert et al. 2017). From this perspective, partisanship is a social identity that remains stable even as leaders and platforms change. Expressive partisanship involves motivated political reasoning in defense of the party, encourages the vilification of threatening out-parties, and leads to action-oriented emotions that result in heightened political activity. Most importantly, and at odds with the traditional Rational Choice approach, these cumulative processes minimize strong partisans' reactivity to accusations of poor party performance, weak leadership, or an altered platform resulting in a relatively stable political identity (Green et al. 2002). Moreover, partisan identity is likely to strengthen over time as a young voter consistently supports one party over others in successive elections (Dalton and Weldon 2007). The expressive approach to partisanship is grounded in social identity theory (Green et al. 2002; Huddy et al. 2015).


    Social identity theory has been confirmed in numerous studies and the vast body of research that originated from it lends considerable insight into the dynamics of social and partisan identities (Huddy 2013). This research reveals that the effects of a social identity are most pronounced among strong identifiers, underscoring the need to reliably measure gradations in identity strength (Huddy 2001, 2013), which led to the creation of the new multi-item measure of partisan identity strength (Huddy et al. 2015).



    Item generation and selection

    The partisan identity scale is based on the Identification with a Psychological Group (IDPG) scale created by Mael and Tetrick (1992). We adapted five items from the original scale, and added three new items. We selected these items to capture a subjective sense of group belonging, the affective importance of group membership, and the affective consequences of lowered group status, which are crucial social identity ingredients (Ellemers et al. 1999; Leach et al. 2008).



    The partisan identity scale was included in nationally representative surveys in three different countries. The scale is assigned to partisans (i.e. voters with feel attached or feel closer to a certain political party.). The items were presented in a random order. Women in all three samples tend to have lower values on the partisan identity scale than their male counterparts with an average difference of 0.3 on a 0-1 scale (p < 0.001), as do younger voters compared to older voters (p < 0.001).


    The Netherlands

    We employ data collected before and after the 2012 Dutch Parliamentary elections among members of the Longitudinal Internet Studies for the Social Sciences (LISS) panel. The LISS contains 5000 households, entailing 8000 individuals, drawn as a true probability sample of households in the national population register maintained by Statistics Netherlands. Data are drawn from three time points: prior to the national election in August 2012 ("Elections 2012"), after the national election in September 2012 ("Dutch Parliamentary Election Study"), and again as part of a module in December 2012/January 2013 ("Politics and Values: Wave 6"). The partisan identity scale was included in the August 2012 pre-election wave. Data in all three waves were used to validate the scale. We examine 4680 respondents who were identified as having a partisan preference and completed the identity items in the pre-election module.



    Swedish data were drawn from the Swedish Citizen Panel, a largely opt-in online panel run by the Laboratory of Opinion Research (LORE) at the University of Gothenburg. We utilize data from Panel 8 (11/14/13-12/18/13) and add-on Panel 8-2 (12/10/13-1/7/14). 16,130 panelists were invited to take the Panel 8 survey and 9279 completed it for a completion rate of 64 %. 2000 panelists were invited to complete Panel 8-2 of which 1496 answered the survey. All panelists in Panel 8.2 and a randomly selected set of 2000 panelists in Citizen Panel 8 received the identity model. Our sample is confined to those in Panel 8 and Panel 8-2 who completed the identity items (N = 2464).


    United Kingdom

    Data for the U.K were taken from the 2015 British Election Study (BES), an online panel study conducted by YouGov. We draw on data from pre-election wave 3 of the BES, conducted between September 19, 2014 and October 17, 2014 and pre- election wave 4, conducted in March 2015. In total, 27,839 respondents participated in wave 3 and 6141 were randomly assigned and 5954 completed a module that included the partisan identity items. In wave 4, 16,629 respondents participated and 3500 of them completed the partisan identity module. The more numerous data from wave 3 are used to examine the measurement properties of the partisan identity.


    Item analyses

    We use an Item Response Model to demonstrate that the eight items supplement each other to cover a broad range of the underlying partisan identity trait in each country. Since each item in the partisan identity scale contains four response categories (e.g., rarely/never, sometimes, often and always in the Netherlands and Sweden), we apply a Non-Rasch Model for ordered polytomous data, namely the Graded Response Model (Samejima 1974).[1] Based on the information provided by each item's response categories, we generated an information function for each item and then for the scale as a whole for each country. Figures 1, 2, and 3 contain the graphical representation of each item information function by country. Graphs were created using the ltm package (Rizopoulos 2006) in R.  


    The individual items vary in amount of information as well as in their ability to capture high or low levels of partisan identity. In all three countries, two items display multiple peaks below the midpoint of the latent trait continuum and thus, provide especially good coverage of lower levels of partisan identity: "When I meet someone who supports this party, I feel connected", indicated in purple, and "When people praise this party it makes me feel good" indicated in black. A third item, "I have a lot in common with other supporters of this party," indicated in dark blue, also provides reasonable information at lower levels of partisan identity. In contrast, the item "When I speak about them, I refer to them as my party" provides good coverage of higher levels of partisan identity in all three countries.




    Figure A. Partisan Identity Item Information Functions (Netherlands).



    Figure 2. Partisan Identity Item Information Functions (Sweden).





    Figure 3. Partisan Identity Item Information Functions (United Kingdom).


    Item parameter

    Based on an unconstrained graded response model, we calculated the total information provided by each item in the scale, the overall scale, and the short scale (comprised of items with a *) (see Table 1). Items that provide high levels of information measure the underlying latent trait (i.e. partisan identity strength) with a high level of precision.


    Table 1

    Total information provided by each item in the scale, the overall scale, and the short scale


    Total Information




    United Kingdom

    When I speak about this party, I usually say "we" instead of "they".




    I am interested in what other people think about this party.




    When people criticize this party, it feels like a personal insult. *




    I have a lot in common with other supporters of this party.




    If this party does badly in opinion polls, my day is ruined.




    When I meet someone who supports this party, I feel connected with this person. *




    When I speak about this party, I refer to them as "my party". *




    When people praise this party, it makes me feel good. *




    Partisan Identity Scale




    Short Partisan Identity Scale




    Note. * Items of the short scale.



    [1] The Graded Response Model is an extension of the two-parameter logistic (2PL) model for graded response data in the sense that it allows the discrimination parameter α to vary across items. 


    Standardized instructions, evaluation rules, and preliminary reference values guarantee the objectivity of the scale in execution, evaluation and interpretation.



    We calculated Cronbach's Alpha for each sample. The scale shows a high level of reliability across all three countries. Numbers in parentheses indicate the reliability of the 4-item version of the scale.


    Netherlands Alpha = 0.86 (0.79)

    Sweden Alpha = 0.83 (0.78)

    United Kingdom Alpha = 0.88 (0.86)



    We assessed the scale's predictive validity and show that the partisan identity scale is a stronger predictor of in-party vote and political participation than the traditional partisan strength item, a measure of ideological intensity, or other conventional predictors such as education. This is true for both the full and the short scale: Partisan identity strength has a sizeable influence on in-party voting and its effects exceed that of partisanship strength. The probability of voting for one's party ranges from a low of roughly .45 in the Netherlands and .50 in the U.K. at the lowest levels of partisan identity to a high of .90 for those at the highest levels. In contrast, the probability of in-party voting changes less dramatically across the range of partisanship strength, ranging from a low of .40 to a high of .75 across the range of strength in the Netherlands, and just under .60 to above.80 in the U.K. The effect of both partisanship and identity strength on in-party voting is reduced in Sweden where partisan loyalty was high across the board.


    The analysis of political participation is similar to that of in-party voting: Both partisan identity and partisanship strength are significant predictors of political engagement, although the effects of partisan identity are roughly twice as large as those of partisanship strength when entered into separate equations, and over four times as large when both are entered simultaneously: Partisan identity has greater influence on participation in Sweden and lesser influence in the Netherlands.  For example, in Sweden, participation ranges from .37 at the lowest level of partisan identity to .75 at the highest. In contrast, the effects of partisanship strength are more muted (ranging from .36 to .57 across the range of partisanship strength). For more information, see Bankert et al. (2017).


    Descriptive statistics (scaling)

    We calculated the means and the standard deviations of the partisan identity scale for each country's sample (see Table 2). Mean estimates are based on the original 1-4 scaling. The samples meet the requirements for a standardization sample.


    Table 2

    Means, Standard Deviations, Skewness, Kurtosis of Scale Scores



    Standard Deviation













    United Kingdom





    Note. Scale ranging from 1 (never/rarely) to 4 (always).


    Further quality criteria

    We tested for measurement invariance of the scale using multiple group confirmatory factor analysis (MGCFA) for ordered categorical data using the lavaan package in R (see in detail Bankert et al. 2017). The scale has been shown to be metric invariant across countries but failed to meet cutoff criteria for scalar invariance. However, since results of traditional invariance tests often rely on arbitrary cutoff lines, sensitivity analyses were used to further examine scalar invariance. Following Oberski (2014), we examine invariance sensitivity, "...the likely impact of measurement differences on substantive comparisons of interest". (p.3). We compute the EPC-interest, which is a measure of the expected change in the parameter of interest, partisan identity in this case, when freeing a particular equality constraint. With the EPC-interest we can evaluate whether it is feasible to compare partisan identity means across countries by estimating the change in partisan identity if certain invariance restrictions (such as equivalence constraints on a scale item's intercept) are removed. Put differently, the EPC-interest evaluates directly whether a violation of measurement invariance also leads to biased estimates of partisan identity in different countries (Oberski 2014).


    Our analyses reveal an EPC-interest at or below .076 in absolute value. This number is significantly smaller than the latent mean differences of partisan identity across countries (the absolute difference between the U.K and Netherlands is .133, the U.K. and Sweden is .645, and the Netherlands and Sweden is .512). The minor influence on the latent party identity trait scores indicates that our substantive conclusions regarding the comparison of partisan identity across countries are not changed to any great degree by potential model misspecifications. Even when the requirements of the scalar invariance model are relaxed the magnitude of partisan identity remains relatively constant. Overall, these results provide evidence that the partisan identity scale exhibits features of scalar invariance, and works similarly in all three countries. For more, see Bankert et al. (2017).


    Like any measurement of self-report, it is possible participants misreport partisan identity strength because of social desirability bias or priming of partisanship or even acquiescence bias. However, given the consistently strong effects of partisan identity on political behavior, these potential measurement issues seem to be minor. The scale is also very economical with a processing time of less than two minutes.


    Further literature

    Bankert, A., Huddy, L., & Rosema, M. (2017). Measuring Partisanship as a Social Identity in Multi-Party Systems. Political Behavior39(1), 103-132.


    Alexa Bankert, Department of Political Science, School of Public and International Affairs, The University of Georgia, E-Mail:

    Boije, Edvin & Dahlberg, Stefan (2014) The 2014 Internet Campaign Panel. Swedish National Election Studies. University of Gothenburg.

    Fieldhouse, E., J. Green., G. Evans., H. Schmitt, C. van der Eijk and J. Mellon (2014) British Election Study Internet Panel Wave 3.

    Van Ham, Carolien, Aarts, Kees (2012): Dutch Parliamentary Election Studies 2012.