Who we are
Halstead Research Group LLC was formed in 2019 to do one thing well: build and look after a consumer panel that research teams can field studies against without having to construct recruitment machinery of their own. Recruiting the right households is slow, unglamorous work. It involves writing screening questions that people answer honestly, reviewing applications by hand, keeping records clean, removing duplicates, and contacting members rarely enough that they still read the email when it arrives. Most research teams do not want to own that work. We do.
Shopper Panel USA is the public face of the panel. If you apply through this site, if you are invited to a study, or if you write to us with a question, you are dealing with the Shopper Panel USA panel team, and the company standing behind that team is Halstead Research Group LLC. The two names refer to the same operation. We keep them distinct on paper because the legal entity may serve more than one research program over time, while the panel a member joins should have one stable identity for as long as they are on it.
We are deliberately a small operation. Panel quality is a function of how carefully members are screened and how sparingly they are contacted. It is not a function of how many addresses sit on a list, and we do not treat it as one.
What a consumer research panel is for
A great deal of what passes for consumer research is an open survey link that anyone can click. That method has a structural problem: the people who answer are the people who happened to see the link and had a reason to click it. A survey about laundry detergent circulated openly will be answered disproportionately by people with strong opinions about laundry detergent, and the result reads as a finding when it is really an artifact of who showed up.
A panel exists to remove that problem. Because the members are known and screened before any study is written, a study can be fielded to a defined composition of households rather than to whoever arrives. If a research design calls for a read across four household sizes and five purchase-frequency bands, the panel can be sampled to fill those cells, and the analyst can say what the sample was rather than guessing at it.
A standing panel also makes two things possible that a one-off link cannot do. The first is repeat measurement: asking the same households the same questions in March and again in September produces a change over time in the same people, which is a far stronger signal than two separate surveys of two separate crowds. The second is de-duplication. Because each member holds one membership tied to one verified email address, the same person cannot answer a study five times, and a household cannot be counted as five households.
The value of a panel to a research team, in short, is that it turns "we asked some people" into "we asked this defined group, and here is who they were."
How we work with research teams
Studies reach the panel through a fixed sequence. It exists so that no study is fielded before it is clear what question it answers and who should be asked.
1. Brief and research design
A research team brings us a question and the decision that hangs on it. We work out with them what kind of instrument answers that question honestly. Some questions are best answered by a short structured survey; some need a diary kept over several weeks; some need a conversation. Choosing the wrong instrument is the most common way a study produces confident nonsense, so this step is not a formality.
2. Sampling plan and quotas
We translate the design into a sampling plan: which screening profiles are eligible, how many completions are needed in each cell, and which profiles must be excluded to keep the read clean. Quotas are set before invitations go out, never adjusted afterwards to make a result look tidier.
3. Invitation
We invite only the members whose stored screening answers match the plan. An invitation states the topic in plain language, what taking part involves, roughly how long it takes, and what happens to the answers. There is no obligation attached to it. Members who ignore an invitation are not penalized and are not contacted again about that study.
4. Fielding
The study runs for a stated window. We monitor which quota cells are filling and which are not, and invite further matching members if a cell is short. If a cell cannot be filled honestly, we report it as unfilled rather than relaxing the definition to reach a number.
5. Quality review
Before anything is delivered, responses are reviewed for quality: completion times far below what the instrument allows, straight-lined answer patterns, contradictory answers between screening and study, and open-text responses that appear to be machine-generated. Responses that fail review are removed from the analysis set, and the count of removals is reported alongside the findings rather than hidden.
6. Delivery
Findings go to the client that commissioned the study, ordinarily in aggregated or de-identified form, together with the sampling plan, the achieved sample composition, the field dates, and the quality exclusions. A finding without that context is not a finding, and we do not deliver one.
Our stance on data minimization
We ask for the least information that lets us do the work, and we ask for it as late as we can.
The application on this site collects five screening answers, a first and last name, and an email address. That is the whole record we create at application time. We do not ask for a street address, a date of birth, an income figure, an employer, a phone number, or any account credential, because none of those are needed to decide whether a household matches an open study. Anything a specific study genuinely requires beyond that is asked inside that study, is explained in that study's own notice, and is asked only of the people taking part in it.
There are categories of information we actively do not want. We do not ask for payment card or bank details, Social Security numbers, driver's license numbers, passwords, or health records, and we have no process that would use them. If someone sends information of that kind to us anyway, we delete it. This is worth stating plainly because it is also the single most reliable way to identify a message that is not from us: a request from "the panel" for your bank details or your password is not ours.
Retention is bounded rather than indefinite. Applications that are not accepted are held for a limited period so that a later study can be matched against them, and then deleted. Member records are deleted or de-identified after a defined period of inactivity. The exact periods are set out in section 7 of the Privacy Policy, and a member can ask us to delete a record sooner than any of them.
Consent and participant rights
Consent on this panel is given per study, not once and forever at the point of joining. Accepting membership means agreeing to be told about studies. It does not commit anyone to take part in any of them.
Every invitation carries the same four facts before a member decides: what the study is about, what taking part involves, approximately how long it takes, and what is done with the answers. Where a study needs identified rather than de-identified responses, that is stated in the invitation, not buried in a notice afterwards, and a member can decline on that basis alone. Declining has no effect on membership and no effect on future invitations. Silence is treated as a decline.
Members can ask us at any time for a copy of what we hold, correct anything inaccurate, or have the record deleted. Leaving the panel takes one line of email; we close the record and confirm in writing. None of these requests require a reason and none of them are met with a retention offer.
Why membership is by application
Open sign-up produces the wrong panel. It fills quickly with people who join every panel they encounter, it produces duplicate memberships across households, and it skews the standing pool toward frequent survey takers, whose answers stop resembling the general population fairly quickly. A panel assembled that way is easy to build and worth very little.
Applying is therefore a filter with a purpose. The five screening questions tell us the household profile; the review tells us whether that profile is one that current and forthcoming studies actually need. Applications are reviewed on a weekly cycle by a person, not by an automated score.
The consequence is that many well-qualified applicants are not invited, and that this is normal rather than a judgment about the applicant. Whether a profile is needed depends entirely on what research is open at the time. A profile that no study calls for in one month may be exactly the profile a study needs in the next, which is why applications are kept on file and matched against studies as they open.
What we do not do
- We do not sell member contact details, and we do not supply them to advertisers or list brokers.
- We do not run advertising trackers, analytics pixels, or third-party marketing scripts on the pages of this site.
- We do not ask members for payment details, passwords, or government identification numbers, for any reason, at any point.
- We do not publish research that identifies an individual member by name without that member's separate written permission.
- We do not adjust a sampling plan after fielding to make a result more attractive to the client that commissioned it.
Firm details
- Legal entity
- Halstead Research Group LLC
- Panel operated
- Shopper Panel USA
- Founded
- 2019
- Registered office
- 1309 Coffeen Avenue, Suite 1200
Sheridan, WY 82801
United States - Correspondence
- support@shopperpanelusa.com, or by mail to the address above. We do not operate a telephone line.
- Panel scope
- Adults aged 18 or older resident in the United States. One membership per person.