The reports, comprised of information culled from various Internet sources, were riddled with inaccuracies, and several included very little information, the authors said.

The errors ranged from mundane reporting of inaccurate phone numbers or email addresses to more serious errors and omissions about the subjects' education and income.

Seven of the 15 reports from eBureau, which touts its ability to predict income through its advanced models and offers insights based upon education, contained errors in estimated income, nearly doubling the salary of one participant and cutting the salary of another one in half.

In addition, 11 of the 15 reports incorrectly stated the subjects' education level. Reports purchased from Intelius and Spokeo had the most inaccuracies, while the reports from Acxiom, eBureau, and ID Analytics contained very little information, the researchers said.

Four of the five firms declined to comment or declined to return reporters' telephone calls and email queries for this story. Only Spokeo responded, citing its role as a "people search" firm not qualified to provide data applicable to measuring consumer credit risk.

"Spokeo is not a consumer reporting agency," said Angela Saverice-Rohan, the company's general counsel and chief privacy officer. "Anyone purchasing a Spokeo account must respond to our compliance assurance questionnaire asking about the intended use of the data. Based on the response, we will only allow access that is in accordance with our terms, which expressly prohibits use for any eligibility determination governed by the Fair Credit Reporting Act."

Spokeo's denial of culpability may fall on deaf ears when it comes to the FCRA's standards and guidance, the NCLC authors said. The act's imposition of requirements on consumer reporting agencies are substantial and compliance assurance questionnaires offered by Spokeo or any other big data providers may not be sufficient protection against the data being misused, they said.

"Three of the most important functions of the FCRA deal with accuracy, disclosure, and the right to dispute items on the report," wrote Yu and McLaughlin. "It is highly unlikely, given the size of the data set and the sources of information, that the companies that provide big data analytics and the users of that data are meeting these FCRA obligations."

The problem is further compounded by the estimated 64 million Americans with insufficient or no credit history to facilitate effective credit scoring. In an attempt to fill in the background – and generate more profitable loans – some lenders scour the Internet, social media and mobile apps, incorporating inaccurate information from big data companies as part of consumers' personal credit record.

Since many of these credit-shy individuals come from low-income or disadvantaged households, such data collection methodologies also need to be analyzed for potential discriminatory impact, the NCLC said. Big data firms' use of undisclosed analytics make such analysis difficult if not impossible, the study said.

"By law, the calculation of your credit score cannot use or take into account factors such as race or color, religion, gender, national origin or marital status," according to the CFPB website.

None of the big data companies contacted by NCLC for the survey were listed as approved consumer credit reporting agencies, but that's not likely to keep some lenders from accessing the data to create a weightier even if inaccurate background for potential borrowers, the NCLC said.

Payday lenders and other proponents of big data underwriting argue that by using multiple factors to price credit, the cost of credit will be reduced for low-income borrowers, enabling lenders to provide lower-cost small loans as alternatives to payday loans. NCLC evaluated seven loan products based on big data underwriting. Six of these products present themselves as payday loan alternatives, yet still carry triple-digit annual percentage rates.

What's more, five of the seven require borrowers to provide their bank name, routing number and account number. A lender could potentially use this information to reach into a bank account and take the funds if the consumer fails to make a payment, thus ensuring that the lender will be repaid prior to life-sustaining necessities, such as rent or food, and trapping borrowers into a cycle of debt, the study said.

As a result of the study's finding, the NCLC made several recommendations to both the Federal Trade Commission and the CFPB. Researchers urged the agencies to study big data brokers, testing for accuracy, discriminatory impact and legal compliance with FCRA and the Equal Credit Opportunity Act.

"Unfortunately, our analysis concludes that big data does not live up to its big promises," the researchers concluded.