One flat data frame of relaxed harmonized variables (experimental)

qes_decon() returns one data frame for every harmonized Quebec Election Study, with one row per respondent and wave and one column per concept, under plain names in the style of the cesR package (education, income_cat, vote_choice, sovereignty, lr...). It is a relaxed harmonization: one concept goes in one column for every study, even when the wording or the answer options differ, with coarse common categories (education in three groups, income in thirds of each study's respondents, interest low, medium or high, a referendum vote of yes or no whatever the question). It trades exactness for coverage: use qes_harmonize() and its grades when the difference between two questions matters.

Usage

qes_decon(studies = NULL, lang = c("en", "fr"), weights = TRUE, quiet = FALSE)

Arguments

studies: NULL (default) or "all": every harmonized study (the 11 Quebec studies; the 1998 firms' own files are in qes1998). Otherwise study codes, as qes_studies() lists them; "qes_demo", the synthetic demonstration study, is accepted and needs no download.
lang: Language of the factor levels, of the label, relaxed and sources attributes and of decon_sources: "en" (default) or "fr". Column names do not change.
weights: If TRUE (default), adds weight and weight_var.
quiet: If TRUE, no download, cache or summary messages.

Value

A data frame (no class of its own), returned visibly, with the columns study, year (the study's year; for the CROP polls, the year of the respondent's poll), wave, qes_id (<study>:<identifier>), then weight and weight_var (with weights = TRUE), then the relaxed columns listed under Columns: factors (ordered for ordinal ones) with every category as a level, or numbers. Each relaxed column carries attr(, "label"), attr(, "relaxed") and attr(, "sources") (a data frame: study, wave, source_var, wording_ref, recode). The data frame carries decon_sources (column, study, wave, base, source_var, wording_ref, recode, relaxed (TRUE where a relaxed mapping or a collapse was used), status and applied of the source row, n_value (respondents of a static column, rows of a wave column) and na_reasons ("dk=12; refused=3")), weights (study, wave, weight_var, status, reason: NA, "needs_review", "no_recommended_weight" or "not_in_data", n_rows, n_weight, mean), qes_no_wave (study, n_rows: the respondents left out because they took part in no wave), qes_provenance (the files read; see qes_provenance()), spec_version, qesR_version and lang.

Details

Each column says how it was relaxed (attr(, "relaxed"), one sentence) and where each study's values come from (attr(, "sources"): the source variable, the document that gives its wording and the rule that turns its codes into the column's categories). attr(x, "decon_sources") gathers them for every column, with the count of values and of each reason for a missing value. Relaxed columns carry no comparability grade, and none of them claims that two studies asked the same question.

A column is built from a strict target or a pooled variable of the harmonization spec where one exists (gender, vote_choice, satis_democracy...), recoded into the column's categories where needed (born_quebec from the birthplace, interest from the 0-1 interest scale), and from relaxed mappings of the studies' own questions where the strict layer has none or keeps a study out (education, income_cat, religion, employment...). The relaxed mappings are reviewed like the strict ones: a mapping not yet signed off (status review) is not applied, and its cells are NA with reason not_reviewed (see qes_spec("relaxed_maps")); a message says how many. The 65 mappings of spec 4.6.0, the 5 it adds (the education of qes1998, union and econ_self) included, are signed off (status stable) by an automated double review against the original files and documents, not a human review, as their reviewed_by and review_note say. A strict row still in review is not applied either: the previous vote (vote_prev) of qes2008 and qes2018 is NA (reason not_reviewed) while a decision on the respondents too young to vote at the previous election is pending, and the summary message names such columns and studies.

Rows, waves and weights

The rows are those of qes_harmonize(layout = "long"): one per respondent and wave they took part in, so a panel respondent has one row per wave, and wave is never NA. A respondent who took part in no wave has no row (in qes2007_panel, one interview that neither wave's rule counts); attr(x, "qes_no_wave") counts them, study by study. Socio-demographic columns (education, income, language, religion, region...) are the same on every row of a respondent; the vote, turnout and attitudes sit on the row of the wave that asked them (the vote intention on the campaign wave, the reported vote on the post-election wave). With weights = TRUE, weight is the wave's recommended weight rescaled to mean 1 in each study and wave, and weight_var its source variable. Both are NA where the study has no recommended weight (qes2008: its weights are calibrated on the reported vote) or where the weight still needs review; attr(x, "weights") says which, study by study and wave by wave (its n_rows add up to the rows of x). Estimate within one study and wave, or with qes_design() on qes_harmonize() output.

Who was surveyed

These differences between the studies matter for most analyses.

qes1998 (the CROP and CREATEC polls) surveyed francophones only, each firm by its own definition (mother tongue for CREATEC, the language spoken most often at home for CROP): language is French or NA there, and the study says nothing about non-francophones.
qes2018 also interviewed people aged 16 and 17 (251 of its 3,072 respondents, by the deposit's note). They are NA in age_group and vote_choice but have values in the other columns (yob, gender, sovereignty...): drop them (by yob) to describe the electorate.
In qes2018_panel only the web respondents were asked about employment: its 400 telephone respondents appear in employment as don't know or refused (NA, reason dk_refused), not as a real answer.
religion comes from two formats: 2012 to 2018 first ask whether the respondent belongs to a religion, then which one; 2022 offers one long list (agnostic counts as no religion). The share with no religion is higher after a filter question, so the levels are not strictly comparable across that change.

Correspondence with cesR

cesR's get_decon() returns 21 columns of the 2019 Canadian Election Study. Their counterparts here: citizenship, yob, gender, education, lr (cesR's lr, lr_bef and lr_aft: no study asks the scale both before and after the election), religion, language, language_eng and language_fr (no study asks about Indigenous languages), employment, income_cat (the income amount is not comparable across studies), marital, econ_retro (Quebec's economy) and econ_self (personal finances, asked in qes2022 only, kept by the owner's decision of 2026-10-01). province_territory is dropped (every respondent lives in Quebec), and so are sexual orientation and the federal economy (one study at most). union (union membership, three studies, kept by the same decision) and education4 (the four groups of education that education joins into three) have no counterpart among cesR's columns.

Columns

The relaxed columns of spec 4.7.0, in order (qes_spec("relaxed") lists them with their French labels):

citizenship: Citizenship (one value per respondent, target:citizen). Canadian citizenship where a study recorded it: asked directly in 2022, and taken in the 2018 panel from the screening question on the right to vote in the coming Quebec election, so that every respondent of that panel is a citizen. Levels: Canadian citizen; Not a Canadian citizen.
yob: Year of birth (one value per respondent, target:birth_year). The year of birth as reported; the studies that asked only an age group have none (see age_group). A number from 1900 to 2010.
age_group: Age group (one value per respondent, target:age_group3). Three age groups (18-34, 35-54, 55 and over), from the study's own age bands or from the age at the start of fieldwork; respondents under 18 are missing. Levels: 18-34; 35-54; 55 and over.
gender: Gender (one value per respondent, target:gender). Gender or sex as each study asked it: man or woman everywhere, and in 2022 also non-binary or another gender; the questions differ, the categories do not. Levels: Man; Woman; Non-binary; Another gender.
education4: Education (four groups) (one value per respondent, relaxed mappings only). Each study's levels are grouped into four: no high school diploma, high school, college (CEGEP, technical or trade school) and university, including some university; the 2007 panel and the CROP polls asked years of schooling, so their high school group also holds those who left secondary school without a diploma, and their college group may hold some respondents with some university but no degree. qes1998 is missing (not mappable): the middle group of its pooled file, 10 to 15 years of schooling, spans high school and college (education, in three groups, includes it). The strict target of the same name in qes_harmonize() groups differently: its lowest group is primary or less. Levels: No high school diploma; High school diploma; College, CEGEP or trade school; University.
education: Education (one value per respondent, column:education4). Each study's levels are grouped into three: no high school diploma, high school to college (a high school diploma, CEGEP, technical or trade school) and university, including some university. These are the groups of education4 with high school and college joined, which lets qes1998 in. Where a study asked years of schooling (the 2007 panel, the CROP polls and the CROP respondents of qes1998), the bands do not follow diplomas: some respondents without a diploma (10 years in qes1998; 8 to 10 years in the others, as Quebec high school ends after 11 years) count as high school to college, and some university students (14 or 15 years) are not counted in university. Levels: No high school diploma; High school to college (CEGEP); University.
income_cat: Household income (thirds) (one value per respondent, relaxed mappings only). Household income in thirds of each study's own respondents: income brackets are never split, so a third holds the whole brackets whose midpoint falls in it and is rarely exactly a third; the dollar limits differ from study to study. Levels: Low (bottom third); Middle; High (top third).
language: Mother tongue (one value per respondent, target:lang_mother). The language first learned in childhood, in three groups; a respondent who reported French and another language is French, and English and a language other than French is English, and the 1998 CREATEC respondents are French by the design of their sample. Levels: French; English; Other.
language_fr: French as a mother tongue (one value per respondent, column:language). Yes when French is among the mother tongues reported, so a respondent with two mother tongues can be yes in both language_fr and language_eng. Levels: Yes; No.
language_eng: English as a mother tongue (one value per respondent, column:language). Yes when English is among the mother tongues reported, so a respondent with two mother tongues can be yes in both language_eng and language_fr. Levels: Yes; No.
religion: Religion (one value per respondent, relaxed mappings only). The religion the respondent belongs to, in five groups; the 2022 question offers a long list, where agnostic counts as no religion, and the other studies first ask whether the respondent belongs to a religion at all. Levels: Catholic; Protestant; Other Christian; Other religion; No religion.
marital: Marital status (one value per respondent, relaxed mappings only). Married or living with a partner, separated or divorced, widowed, or never married; 2012 and 2014 asked the official civil status, with no common-law option, so some partners who live together are never married there. Levels: Married or living with a partner; Separated or divorced; Widowed; Single, never married.
employment: Employment (one value per respondent, relaxed mappings only). The main employment status in five groups; a respondent who gave two statuses, such as retired and working, takes the one that is not work, and at home, unable to work and other statuses are other. In the 2018 panel only web respondents were asked about employment, so its telephone respondents appear as don't know or refused. Levels: Working (employee or self-employed); Unemployed; Retired; Student; At home, unable to work or other.
union: Union membership (one value per respondent, relaxed mappings only). Whether the respondent belongs to a union in 2022, but whether the respondent or anyone in the household does in 2014 and 2018 (in 2018, respondents who live with their parents were asked about their family: parents, brothers or sisters). Only these three studies ask it, and 2022 asks it after the election, so its campaign-only respondents are missing. Levels: Yes; No.
region: Region (one value per respondent, target:region_cma3). The Montreal census metropolitan area, the Quebec City census metropolitan area or the rest of Quebec, from each study's region or sub-region variable, with the boundaries each study used. Levels: Montreal CMA; Quebec CMA; Rest of Quebec.
region_admin: Administrative region (one value per respondent, relaxed mappings only). The 17 administrative regions of Quebec, where a study recorded them or sub-regions that fit within them; the sub-regions of the 2007 study and the 2007 panel that split a region are joined. Levels: Bas-Saint-Laurent; Saguenay–Lac-Saint-Jean; Capitale-Nationale; Mauricie; Estrie; Montréal; Outaouais; Abitibi-Témiscamingue; Côte-Nord; Nord-du-Québec; Gaspésie–Îles-de-la-Madeleine; Chaudière-Appalaches; Laval; Lanaudière; Laurentides; Montérégie; Centre-du-Québec.
born_canada: Born in Canada (one value per respondent, target:born_canada). Born in Canada or not, as each study asked it: from the birthplace (Quebec, elsewhere in Canada or abroad), or asked directly in 2022. Levels: Yes; No.
born_quebec: Born in Quebec (one value per respondent, target:birthplace3). Born in Quebec or not, from the birthplace question (Quebec, elsewhere in Canada or abroad); 2022 asked only whether the respondent was born in Canada, so it is missing there. Levels: Yes; No.
vote_choice: Provincial vote choice (on the wave that asked it, pooled:vote_choice). The reported vote where a study asked it, else the vote intention with the undecided pushed toward the party they lean to, else the first intention question (the pooled vote_choice); vote_type says which, and the parties are those each study offered. Levels: PLQ; PQ; CAQ; QS; PVQ; PCQ; ON; ADQ; Other party; Would not vote / none / would spoil.
vote_type: Question of the vote choice (on the wave that asked it, pooled:vote_choice__type). Which question vote_choice comes from on each row: the reported vote, the pushed intention or the first intention question. Levels: Reported vote (after the election); Vote intention (undecided pushed); Vote intention (first question).
turnout: Turnout (reported) (on the wave that asked it, pooled:turnout). Whether the respondent says they voted in the Quebec general election of the study, asked after it; the wordings and answer options differ, some offering several ways of not voting. Levels: Yes; No.
vote_prev: Vote at the previous provincial election (on the wave that asked it, target:vote_prov_prev). The party of the reported vote at the Quebec general election before the study; those who did not vote are missing, the sources name the election recalled, and the 1998 question names only the PLQ and the PQ. Levels: PLQ; PQ; CAQ; QS; PVQ; PCQ; ON; ADQ; Other party.
pid: Provincial party identification (on the wave that asked it, target:pid_prov). The Quebec party the respondent identifies with, or none, as each study asked it; the parties offered differ from study to study. Levels: PLQ; PQ; CAQ; QS; PVQ; PCQ; ON; ADQ; Other party; None of these.
lr: Left-right self-placement (0-10) (on the wave that asked it, target:lr_self). Self-placement on a left-right scale from 0 (left) to 10 (right); a scale of another length would be rescaled to 0-10, and every study that asked one used 0 to 10. A number from 0 to 10.
interest: Interest in politics (on the wave that asked it, pooled:pol_interest). Interest on 0 to 1 (interest_01) in three bands, low below 0.55 and high from 0.85: on a 0-10 question, 0 to 5 is low, 6 to 8 medium and 9 or 10 high; on a 4-point question, not at all or hardly interested is low, quite interested medium and very interested high. The two formats still do not line up exactly, and the 2008 study and the 2007 panel asked interest in the election or the campaign, not in politics. Levels: Low; Medium; High.
interest_01: Interest in politics (0-1) (on the wave that asked it, pooled:pol_interest). Interest on 0 to 1 (the pooled pol_interest): four-point answers scored 1, 0.7, 0.3 and 0, and 0-10 answers divided by 10. A number from 0 to 1.
sovereignty: Sovereignty referendum vote (on the wave that asked it, pooled:sov_support). Yes or no in a referendum on Quebec sovereignty, whatever the question (an independent country, a sovereign country, the 1995 question or being favourable to independence); sovereignty_type says which, and would not vote is missing. Levels: Yes; No.
sovereignty_type: Question of the sovereignty vote (on the wave that asked it, pooled:sov_support__type). Which question sovereignty comes from on each row: an independent country, a sovereign country, the 1995 question, being favourable to independence, or the CROP polls' question, whose full wording was not deposited. Levels: Independent country; Sovereign country; 1995 question (sovereignty-partnership); Favourable to independence; CROP referendum question (wording not deposited).
satis_democracy: Satisfaction with democracy in Quebec (on the wave that asked it, target:satis_demo_qc). Satisfaction with the way democracy works in Quebec, on four points, as each study asked it. Levels: Very satisfied; Fairly satisfied; Not very satisfied; Not at all satisfied.
gov_satisfaction: Satisfaction with the Quebec government (on the wave that asked it, target:gov_satisfaction). Satisfaction with the Quebec government of the day, on four points; the 1998 question asks about the Bouchard government, in its own words. Levels: Very satisfied; Fairly satisfied; Not very satisfied; Not at all satisfied.
econ_retro: Quebec's economy over the past year (on the wave that asked it, target:econ_retro_qc). Whether Quebec's economy got better, stayed about the same or got worse over the past year, as each study asked it. Levels: Better; About the same; Worse.
econ_self: Personal financial situation over the past year (on the wave that asked it, relaxed mappings only). Whether the respondent's own financial situation got better, stayed about the same or got worse over the past year. Only qes2022 asks it, during the campaign; the column is kept for analyses within that study (essential), as cesR's econ_self, although it has fewer than three studies. The 2012 and 2014 question on one's own finances if Quebec became independent is another question and is not used. Levels: Better; About the same; Worse.
identity: Québécois and Canadian identity (on the wave that asked it, target:identity_qc_ca). Québécois only, Québécois first, both equally, Canadian first or Canadian only, from one question or from the two orders of a split ballot. Levels: Québécois only; Québécois first, then Canadian; Equally Québécois and Canadian; Canadian first, then Québécois; Canadian only; Other.
attach_quebec: Attachment to Quebec (on the wave that asked it, target:attach_qc). Attachment to Quebec on four points, as each study asked it. Levels: Very attached; Fairly attached; Not very attached; Not at all attached.
attach_canada: Attachment to Canada (on the wave that asked it, target:attach_ca). Attachment to Canada on four points, as each study asked it. Levels: Very attached; Fairly attached; Not very attached; Not at all attached.

Experimental

The relaxed layer is new in qesR 0.9.0 (spec 4.4.0; its mappings signed off in spec 4.5.0). Its columns, groups and mappings may change, and a new relaxed mapping starts in review: qesR 0.9.1 (spec 4.6.0) puts education in three groups instead of four, so that qes1998 can be included, keeps the four groups as education4, and adds union and econ_self. qesR 0.9.2 (spec 4.7.0) cuts interest at 0.55 and 0.85 instead of 0.35 and 0.75, so that a 0-10 answer of 6 to 8 is medium and 9 or 10 high, as "quite" and "very" interested are on the 4-point questions.

En français

qes_decon() renvoie un seul tableau pour toutes les études électorales québécoises harmonisées, une ligne par personne et par vague, une colonne par concept, sous des noms simples à la manière du package cesR. C'est une harmonisation souple : un concept va dans une seule colonne pour chaque étude, même quand le libellé ou les choix de réponse diffèrent, avec des catégories communes larges (la scolarité en trois groupes, le revenu en tiers des répondants de chaque étude, l'intérêt faible, moyen ou élevé, un vote référendaire oui ou non quelle que soit la question). Elle échange l'exactitude contre la couverture : utilisez qes_harmonize() et ses niveaux de comparabilité quand la différence entre deux questions compte. Chaque colonne dit comment elle a été assouplie (attr(, "relaxed")) et d'où viennent les valeurs de chaque étude (attr(, "sources")) ; attr(x, "decon_sources") les réunit toutes. Les colonnes souples n'ont aucun niveau de comparabilité. Les appariements souples sont révisés comme les autres : un appariement pas encore approuvé (statut review) n'est pas appliqué, et ses cellules valent NA (motif not_reviewed). Des 65 appariements de la spécification 4.6.0, y compris les 5 qu'elle ajoute (la scolarité de qes1998, union et econ_self), sont approuvés (statut stable) par une double révision automatisée sur les fichiers et les documents originaux, et non par une révision humaine, comme le disent leurs champs reviewed_by et review_note. Depuis qesR 0.9.1, education compte trois groupes (sans diplôme d'études secondaires, secondaire au collégial, université), ce qui permet d'inclure qes1998, et education4 garde les quatre groupes. Les lignes suivent celles de qes_harmonize(layout = "long") : wave ne vaut jamais NA, et une personne qui n'a participé à aucune vague n'a pas de ligne (attr(x, "qes_no_wave") les compte). Une ligne stricte encore en révision n'est pas appliquée : le vote précédent (vote_prev) de qes2008 et de qes2018 vaut NA (motif not_reviewed), et le message de synthèse le dit. qes1998 n'a interrogé que des francophones ; qes2018 a aussi interrogé des personnes de 16 et 17 ans, NA dans age_group et vote_choice mais présentes dans les autres colonnes ; dans qes2018_panel, seuls les répondants web ont été interrogés sur leur emploi, et ses répondants téléphoniques apparaissent dans employment comme « ne sait pas ou refus » ; la religion vient d'une question filtre (2012 à 2018) ou d'une seule longue liste (2022), si bien que ses niveaux ne sont pas strictement comparables entre ces deux formats. Depuis qesR 0.9.2, interest est coupé à 0,55 et 0,85 : de 0 à 10, 6 à 8 est moyen et 9 ou 10 élevé. Les noms de colonnes restent en anglais ; lang = "fr" donne les étiquettes, les règles et les sources en français.

Examples

# the synthetic demonstration study, offline
d <- qes_decon("qes_demo", quiet = TRUE)
head(d[, c("study", "wave", "gender", "age_group", "vote_choice", "sovereignty")])
#>      study wave gender   age_group vote_choice sovereignty
#> 1 qes_demo post  Woman 55 and over         CAQ          No
#> 2 qes_demo post  Woman 55 and over          QS          No
#> 3 qes_demo post    Man 55 and over         PLQ         Yes
#> 4 qes_demo post    Man 55 and over          PQ         Yes
#> 5 qes_demo post    Man       35-54          QS          No
#> 6 qes_demo post  Woman       35-54         CAQ         Yes
attr(d$sovereignty, "relaxed")
#> [1] "Yes or no in a referendum on Quebec sovereignty, whatever the question (an independent country, a sovereign country, the 1995 question or being favourable to independence); sovereignty_type says which, and would not vote is missing."
attr(d$vote_choice, "sources")
#>      study wave source_var         wording_ref
#> 1 qes_demo post    Q2 / Q3 352010:Q3;352009:Q3
#>                                                                                                                                               recode
#> 1 where Q2: 2 = Did not vote (NA); 9 = Refused (NA); else Q3: 1 = PLQ; 2 = PQ; 3 = CAQ; 4 = QS; 5 = PVQ; 6 = ON; 96 = Other party; 99 = Refused (NA)

# where every column comes from, study by study
src <- attr(d, "decon_sources")
src[, c("column", "study", "source_var", "relaxed", "n_value")]
#>              column    study source_var relaxed n_value
#> 1               yob qes_demo       QAGE   FALSE      60
#> 2         age_group qes_demo       QAGE   FALSE      60
#> 3            gender qes_demo      QSEXE   FALSE      60
#> 4      region_admin qes_demo    QREGION    TRUE      60
#> 5       vote_choice qes_demo    Q2 / Q3   FALSE      44
#> 6         vote_type qes_demo    Q2 / Q3   FALSE      60
#> 7           turnout qes_demo         Q2   FALSE      57
#> 8                lr qes_demo        Q32   FALSE      55
#> 9          interest qes_demo        Q28    TRUE      59
#> 10      interest_01 qes_demo        Q28    TRUE      59
#> 11      sovereignty qes_demo        Q19    TRUE      55
#> 12 sovereignty_type qes_demo        Q19    TRUE      60

# French labels, same column names
d_fr <- qes_decon("qes_demo", lang = "fr", quiet = TRUE)
levels(d_fr$interest)
#> [1] "Faible" "Moyen"  "Élevé" 

# every study: qes_decon() reads the 11 files (downloaded once, then from
# the cache), e.g. d <- qes_decon(); table(d$study, d$sovereignty)