Donors are always asking us “is this a good charity?” There is typically no simple answer to this question. However, Charity Intelligence believes that the best information to help donors with this question is an assessment of the social impact produced by charities for each dollar donated.
Charity Intelligence (Ci) produces consistent, comparable impact ratings for many charities across Canada. These ratings are conservative, evidence-based estimates of the social value that charities create for their clients and the wider community. They are based on two metrics – a Demonstrated Impact Score based on estimates of the charities’ social return on investment (SROI) and a Data Quality Score (DQS). A discussion of each metric can be found below.
Using the Demonstrated Impact Score and the DQS, the Ci team generates an Impact Rating for each charity analyzed. To date we have assessed charities primarily in the social services and education sectors as well as a number in the international aid sector. Ratings appear only for those charities that we have assessed and we continue to add charities to this list, thus more Impact Ratings will appear on charity profiles over time. The Impact Rating appears as a red dot overlaid on a grid:
Impact Rating: Good
Below is the distribution of impact ratings of all charities we have analyzed to date:
Components of the Impact Rating
1) Demonstrated Impact Score (Proven Impact)
We estimate and compare the amount of social good that charities generate per dollar donated. How much good, measured in dollars, do donations accomplish? Social return on investment (SROI) is the best metric we know of for this task because it attempts to measure these amounts directly.
We use standard program evaluation techniques to provide SROI estimates for every major activity by every charity we analyze. We track benefits to both clients and taxpayers/society by estimating the number of outcomes that each charity produces beyond what would have happened absent service. We then multiply these numbers by estimates of the dollar value of each outcome for clients and for society. Benefits to clients include improvements in income, quality of life, and health, while benefits to society consist of increases in tax revenues and public cost savings in areas such as health care, public assistance, and law enforcement. The long-term, discounted values of these benefits are added together and divided by total expenditures to generate a social return on investment/donation (SROI) estimate.
For each charity, we calculate a lower bound, a best estimate, and an upper bound SROI:
- The lower bound SROI is almost entirely based on evidence from the charity, with very few exceptions. It is highly unlikely that the “true” SROI is below this number.
- The best estimate SROI is based primarily on charity data and, where applicable, conservative evidence from external research and/or other charities.
- The upper bound SROI incorporates additional value that the charity could reasonably be producing but that is not yet appropriately backed by evidence.
The Demonstrated Impact Score is a combination of the lower bound, best estimate, and upper bound SROI, with the lower bound and best estimate weighted more heavily than the upper bound. We emphasize benchmark SROI estimates that can be solidly supported by evidence and aim to produce estimates which measure proven impact. This decision to focus on conservative, evidence-supported estimates of results means that better information about a specific charity’s results will typically lead to higher estimates of its demonstrated social impact. This provides an incentive for charities to collect and share better data and diminishes subjectivity in our evaluation process.
To make our SROI estimates comparable across charities, and even across sectors, we regularly examine all causal factor estimates for consistency. As well, each of the inputs used in our model (including outcome values, attribution shares, drop-off rates, and baseline success rates) is based on extensive research. This includes a combination of randomized controlled studies, meta-analyses, and economic cost studies. As we receive more and better evidence, our estimates are regularly updated.
2) Data Quality Score
The second component of the Ci impact ratings is the Data Quality Score (DQS). The DQS measures the quality of a charity’s impact-related evidence. It is calculated as a percentage, using a grading that assesses each charitable program on the data it provides regarding eight main components of SROI: number of unique clients, pre-program client characteristics, program outcomes, counterfactuals, duration of program effects, duration of client engagement, external validation, and spending breakdown. The DQS for each individual charity program is then weighted by the charity’s spending breakdown to determine the overall DQS for the entire charity.
Ci has been measuring the quality of social results reporting for several years through our Donor Accountability Score (previously Social Results Reporting Score). A full explanation of the Donor Accountability rating methodology is available here.
The DQS and the Donor Accountability both measure the quality of information provided by charities. Both ask charities to report a breakdown of their spending by program area, as well as quantified outputs and outcomes that are relevant and timely.
There are, however, two key differences between the Donor Accountability and the DQS:
- Scope:The Donor Accountability has a wider scope than the DQS. It assesses the reporting of a charity’s strategy, activities, outputs, outcomes, learning, and the quality of that reporting. The DQS focuses only on a few key outputs and outcome metrics needed to measure impact, which varies across program types.
- Audience: The Donor Accountability measures how well a charity reports its social results to the general public. In contrast, the DQS measures the quality of the data that the charity collects and shares with our analysts. It does not matter whether this information is published in an annual report or shared privately.