ARCHIVED – An Examination of the Canadian Language Benchmark Data from the Citizenship Language Survey


In the summer of 2006, Citizenship and Immigration Canada requested a comprehensive analysis of an existing data set. Data for the pilot test were collected in six cities from immigrants who were waiting to take their citizenship test. Assessors administered the combined listening and speaking component of the Canadian Language Benchmark Assessment tool (CLBA). In addition, the participants provided demographic information on a wide array of variables. The chief purpose of this research was to examine the relationships between these variables and the language proficiency of the immigrants, as determined by the CLBA scores. The variables of interest were gender, mother tongue, country of origin, length of residence in Canada, nature of language training (LINC or other), full time versus part time language training, length of language training in months, level of formal education, age (chronological), age on arrival, citizenship test score, immigration status and language used at work.

In this report, we will present an overview of the characteristics of the eligible participants in the study, followed by an analysis of the relationships between CLBA scores and participant variables. Additional correlational and regression analyses will then be presented. Problems with the data set will be presented, and finally, recommendations for future research will be suggested.

Definition of ‘Eligible case’

For our study we defined an ‘eligible case’ as an individual whose “mother tongue” was not an official language, who completed the CLBA and who provided consent to participate in the study. There were a total of 3828 cases, later reduced to 3827 that met these criteria (one case was lost in the analysis).

Characteristics of the cases

Slightly more of the participants were females (53.9%) than males (46.1%). The ages of the participants ranged from 18 to 61, with a mean of 35.9 years. Age at immigration ranged from 3 to 56 years, with a mean of 30 years. Length of residence in Canada ranged from 2 to 26 years, with a mean of 5.9 years. According to participants’ reported immigration class, 20% were refugees, 34% were family class, and 46% were independent class.


Only 3 countries contributed more than 5% of the respondents – China (19.3%), India (10.1%), and the Philippines (8.4%) – for a combined total of 37.8%. An additional 20 countries contributed at least 1% of total cases each. See Appendix A for a breakdown by country.


More than 108 mother tongues were represented. A precise number is not possible because of vague classifications or possible misclassifications at the interviews. For example, some individuals were recorded as ‘Chinese’ (vague), ‘Swiss’ (no such language), or Gad (impossible to know whether this was an abbreviation for a language name such as Gaddang). The most frequent 5 mother tongues, each representing more than 5% of the total eligible cases, were Mandarin (15%), Cantonese (7.5%), Tagalog (7.4%), Punjabi (6.5%), and Arabic (5.8%). An additional 18 languages contributed at least 1% of the total cases each. Table 1 presents a breakdown by language grouping (see discussion below) of eligible cases along with counts of English and French speakers.

Table 1: Native speakers of English, French and cases by language group.
Language N Language N
Mandarin 575 Tamil-Dravidian 133
English 513 Korean-Japanese 117
Hindi-Punjabi 350 Serbo-Croatian 106
Semitic 267 Other Slavic 111
Filipino-Indonesian 301 Romance 117
Cantonese 289 Other languages 98
Iranian 260 Vietnamese-Cambodian 59
Somali-Oromo 247 Niger-Congo 49
Other East Indian 223 Turkic 28
Other Chinese 191 French 31
Other European 166    
Russian-Ukrainian 140 Total 4,371

Language training

The cases were almost evenly split between those with language training in Canada (50.7%) and those without (49.3%). Length of training ranged for as long as 114 months, with a median of 6.0 months. Nearly two thirds (61.3%) of those who accessed language training attended full-time, with 38.7% taking part-time classes. With respect to language training progress, only 6.2% indicated that they had completed their language studies, while 67.7% indicated that they were still in the process of studying, and 26.1% reported that they had not completed their language training and were no longer studying an official language.

An examination of Figure 1 reveals that 42.5% of those who received training accessed LINC, another 34.2% obtained fee-based official language training, and 23.3% studied an official language at a high school, college, or university. A Chi square analysis indicated no difference across immigration class in attendance across the three language training groupings
(Chi square [df = 4], 4.22, p = .377).

Figure 1: Source of language training.

Source of language training
Source of language training.
Source Percentage
LINC 42.5
Fee Based 34.2
High School/ College/ University 23.3


Formal education in Canada

The lack of information from the participants about their formal education in their country of origin greatly reduces the usefulness of the information about education in Canada. Half the participants (49.7%) had received formal education in Canada. However, a level was recorded for only 62% of these cases. The data on level of training are therefore also of limited value. It is worth noting that nearly half of those participants who indicated the type of education they had accessed identified studying for a university or college diploma (see Table 2). The next most common response was continuing education, but unfortunately, the diverse array of offerings within this category (including language training) makes it impossible to draw any meaningful conclusions.

Table 2: Formal education in Canada accessed by participants.
Level of Education Percent
University or College Diploma 46.5
Continuing Education 23.5
High School 12.3
Computer Training 11.5
Apprenticeship 6.2

Two thirds of those individuals who had attended formal education programs in Canada were full-time students. Sixty-one percent reported having completed their programs, 26% were still in progress and the remaining 14% identified their status as incomplete.


In order to interpret and analyze the questions on recent and longest held occupations, we recoded the data employing the National Occupation Coding system (NOC). This resulted in 26 categories; the numbers of participants employed in recent occupations are shown in Table 6. Only seven categories accounted for more than 5% of the total participants each. When we compared recent occupations with the longest held occupations, we discovered practically no difference; again, the same seven categories accounted for more than 5% of the total participants. The only difference worth noting was the change in position of Clerical Occupations and Professional Occupations in Natural and Applied Science. In the recent occupation data, 8.3% of the eligible participants reported holding a Clerical Occupation, and 8.1% held a Professional Occupation in Natural and Applied Science. These figures were slightly different in the Longest Held Occupation category, where 8.0% of the respondents were in a Professional Occupation in Natural and Applied Science, and 7.8% held a Clerical position.  Because there were so few differences in these two analyses, we have reported only the recent occupations.

Language at work

More than 85% of participants reported using primarily an official language at work in their most recent occupation: 81.4% of cases reported English, while 4% identified French. The most common nonofficial language category was Chinese: Chinese (3.4%), Cantonese (2.9%), and Mandarin (2.4%). No other language contributed more than 1% of cases to the total. The findings for language used most often at work in the longest held occupation are remarkably similar to the recent occupation percentages.

Citizenship test information

The overall pass rate for the citizenship test was 96.2%. Ninety eight percent of the native English speakers and 100% of the native French speakers passed the test, while 95.9% of the participants with another mother tongue passed. There was no statistically significant difference across the three groups (Chi square [df = 2] = 4.02, p = .134). Nor was there a difference in mean scores on the test, with English speakers scoring an average of 18.8/20, French speakers averaging 19.2 and speakers of other languages averaging 18.5, F (2, 2562) = 2.58, p = .076.

Report a problem or mistake on this page
Please select all that apply:

Thank you for your help!

You will not receive a reply. For enquiries, contact us.

Date modified: