For this step, we want to finalise our decision about what population we want to draw conclusions about and whether we will study every member of the population, or whether we will select a sample and decide how that sample will be selected, and how big that sample will be. These decisions may impact the validity and generalisability of our findings. For instance, if a sample is too small, it would not be valid to generalise our findings to a larger population. Or if we studied only select males, it would not be valid to assume the findings would be the same for females or other genders.

In our projects, we want to either understand the financial abuse or the spiritual abuse of Indigenous elders from a particular community. The sampling technique and the data collection methods used will differ. Let us assume there are roughly 307 elders in that community and there is an alphabetised list of who those elders are. For the financial exploitation study, which we want to study in a quantitative way, we could randomly select 100 of the names, which would be about one-third of the population, and as long as even 78 participated, we would have about 25% of the population included in our sample. What we find or do not find among those randomly selected individuals is likely also to be found or not present among the other elders as well, so our findings can be generalised. When determining your sample size, it is best to have a large sample when possible. A statistical technique known as power analysis can be used to determine the minimum sample size needed to run tests of statistical significance. For the spiritual abuse study, which we want to study in a qualitative way, we would be more interested in choosing participants who we know possess the key characteristics of interest. We would not rely on random sampling techniques; rather, we would strategically select a much smaller number of participants who we would interview using open-ended questions. While we will not be able to generalise in quite the same way as in our quantitative study, our data will be rich and detailed and will likely reveal information we did not anticipate.

The data collection process can take a considerable amount of time. For example, for our project on the financial exploitation of elders, we will cooperate with the tribal social service providers, who will contact the randomly selected elders, arrange a time to visit with them, help them complete the questionnaire and collect it from them. They will also be available to offer services such as a referral to counselling services or information on how to report a crime, if needed and wanted by the elders. For the spiritual abuse example, as mentioned, face-to-face, in-depth interviews will be conducted with a much smaller number of interviewees, but the data obtained will be rich and detailed. These interviews could take hours, scheduling interviews could prove to be challenging, and the analysis of pages and pages of interview notes could take weeks or longer.

Obtaining tribal permission to conduct these studies and then completing the ethical training of the service providers who administer the questionnaires in the financial exploitation study, or conduct in-depth interviews in the spirituality example, also takes time. Several months may be needed to make all the necessary arrangements and receive the required approvals from the tribe as well as our home institutional review board. Indigenous peoples are a protected class under research ethics protocols (see Canada’s Tri-Council Policy Statement), and we will need to demonstrate that our research will benefit, or at least not harm, the participants in our research project. We will need to ensure that the individuals working with the elders understand that an elder can stop taking the survey or interview at any time and that it will not impact the ability of the elder to use the social services normally provided. We will also need to train the service providers to not pry into details about abuse that are revealed in the survey or interview. Then, it will take time to set up the meetings with the elders and collect the responses/conduct the interviews. Collecting one’s own data can be helpful in addressing the initial research question(s), but it can be time consuming and costly to do it well. No matter the method ultimately used and the time this will entail, maintaining ethical standards throughout the research process is vitally important.

Before we set out to collect our data in the next step, we should carefully consider if the variables, hypothesis, and methods are appropriate for exploring the issues we are interested in, if they already exist in a data set we could use, or if we need to gather our own data. There are many accessible and quality sources of social science data. Each of these existing data sets will have some weaknesses, and we cannot change the questions or methods as the project has already been completed, so we have to work with what exists. However, performing [pb_glossary id=”1928″]secondary data[/pb_glossary] analysis is a considerable time- and money-saving strategy. Examples of existing data sets include the UCR2, which illustrates crimes from the perspective of the police; the GSS, which shares the perspective of crime victims and non-victims; and numerous collections of self-report data that reveal the offender’s perspective about crime and its commission. While these existing data sets do not contain the data we need for either of the research questions we are exploring in our illustrative examples in this chapter, it is important to acknowledge the existence of these secondary data sources; as such, these various methods measuring crime in Canada are explored in the textbox below.