How to Start Researching Your Family History: A Step-by-Step Method

What are the first steps in researching your family history?

Most people start in the wrong place: a famous surname, a distant country or a website search box. The reliable start is closer to home, and it costs almost nothing. These seven steps take most families from a box of papers to a working research file in a fortnight.

• 1. Write yourself down first: name, birth date and place, and your source (your own birth certificate).
• 2. Gather what the family holds: certificates, letters, photos, bibles, recordings, and the memories of the oldest relatives.
• 3. Give every person and every item an ID (P01, A01) so notes, scans and records always link up.
• 4. Draw a four-generation chart from memory: ink for what you can prove, pencil for everything else.
• 5. Choose a target you can finish, such as your 14 direct ancestors back to your great-grandparents.
• 6. Write your first research question about one person and one fact.
• 7. Agree a privacy rule with your family: living people decide what is shared about them.

Why should you ask one question at a time?

A topic such as 'the Smith family' produces browser tabs. A question such as 'When did my grandfather arrive in the UK?' produces an answer, or a clear statement that the answer is not in a particular record set. A good question names one person and one fact, and it tells you when to stop: when two independent sources agree, or when you have written down why you prefer one over another.

Before you search, list the record sets that could answer the question for that person's place and period, who holds them and whether they are free, paid, in person or available by letter. Search the cheapest and closest first. This one habit prevents most wasted subscription months.

Where the records are depends on the country and the period. England and Wales have central civil registration certificates, censuses released after a long closure period, and passenger lists for part of the twentieth century. Ireland lost most of its nineteenth-century census returns, so church registers and census substitutes matter more. The United States releases each census after a 72-year restriction and holds births, marriages and deaths at state level. In many African and Asian countries the earliest records for a family are in church, mosque or temple registers, school and court records, newspapers and the memories of elders. Release dates, fees and access rules change, so check the record holder's own guide before you plan around them.

How do you keep a genealogy research log?

A research log is a dated list of every search: who you were looking for, the question, where you searched, which record set and years, the names and spellings you tried, the result, the source found and your next step. The column most beginners skip is the most valuable one: the result, including 'nil'.

A nil result tells you where the answer is not. Six months later, it stops you paying to search the same index again, and it often points to the real problem: the record set does not cover that parish, or the name was written differently. Write one sentence explaining every nil result, and give every next step a follow-up date.

In the Family History Project Kit from Bodhih Training, the Research Log sheet flags any nil result without an explanation and any follow-up that is overdue, and a monthly summary shows hours, searches and finds.

If you want to check how carefully you weigh evidence before you rely on it, AssessAll's Second Look: Research Integrity in Practice assessment is a separate, optional measure. And if you are starting at a new year or a family anniversary, Jobulary's guide to the fresh-start effect explains why such dates help a goal stick.

How do you cite and weigh a family history source?

A citation is the address of a fact. Five parts are enough for a working family historian: who created the record, what it is, when it was made, where in the source the information sits (page, entry, folio, minute), and where you found it, with the date you accessed it.

Weighing a source means asking three questions. Is it an original (the first recording, or an image of it), a derivative (an index or transcript) or an authored account (an obituary or someone's family history)? Was the informant there at the event, or did they hear about it later? Does it answer your question directly, or only indirectly? The Board for Certification of Genealogists sets out a fuller standard, the Genealogical Proof Standard, with five elements: reasonably exhaustive research, complete and accurate citations, analysis and correlation, resolution of conflicting evidence, and a soundly reasoned written conclusion.

A worked example: how big is a four-generation family history?

Each generation back doubles the number of direct ancestors, which is why projects grow faster than people expect. The table uses a composite example from the kit: a researcher in London tracing four generations between Lagos and London over ten months of spare time.

Two numbers matter more than the size of the tree: the share of target ancestors you can name, and the share you have proved with sources. In the example, 12 of 14 are named (86%) but only 5 are proved (36%). That gap is the honest research agenda for the next six months. The same project logged 40 searches in 50.6 hours, 5 of them nil (12.5%), and cited 24 sources with an average score of 3.6 out of 5.

How do you interview elderly relatives about family history?

Interview early: elders hold names, places and reasons that no record captured. Treat the interview as a gift. Read a short consent statement together before recording: the conversation may be recorded, they can pause, skip or stop, they will see the transcript, and nothing about living relatives will be shared without agreement. Record their choices about sharing: family only, a printed book, a reunion, or private notes.

Bring two photos or objects as prompts and ask open questions: 'Tell me about the day you left' works better than 'Did you leave in 1966?' Keep sessions to 45 to 60 minutes, shorter if the person tires. Within a week, send a thank-you and the transcript to check. Copy names and dates into your research log as questions, not facts: a memory is a clue until a record agrees.

How should you look after old family photos and documents?

Originals usually exist once; a scan costs a minute. The Library of Congress advises that relative humidity is the single most important factor in preserving most photographic prints, recommends photo-safe enclosures that meet the relevant ISO standard, handling with clean hands or nitrile gloves, and avoiding 'magnetic' self-adhesive albums. For digital copies it advises keeping at least two copies in places as far apart as practical and checking them at least once a year.

A simple routine works: handle by the edges, scan front and back, name each file starting with its item ID, keep three copies in different places, and store originals somewhere cool, dry, dark and stable. Ask the holder before you publish a photo, and credit them.

Should you take a DNA test for family history?

A consumer DNA test can confirm a suspected relationship, show which side of the family a match belongs to and find living relatives where records are thin. It also reveals information about your relatives, and it can surface unexpected parentage. Decide the question first, then decide who should test: an older relative is often more useful than you.

The US Federal Trade Commission's consumer advice (December 2023) is a good checklist: compare privacy as well as price, check whether the company shares data and lets you delete it, use a strong unique password and multifactor sign-in, and start with the most secure settings. Companies can also change hands, so read what happens to your data if the business is sold. Get informed consent from everyone who tests, record matches with initials, and agree in advance how you will handle a surprise.

What are the most common family history mistakes?

• Copying online trees as if they were proof. Treat them as clues and prove each link.
• Recording facts without sources. Every fact needs an address.
• Throwing away nil results. Log them with a sentence on what they mean.
• Trusting an index without the original record behind it.
• Assuming two people with the same name are the same person. Build a timeline for each.
• Treating ages on records as exact. Turn them into birth-year ranges.
• Waiting too long to interview elders.
• Publishing living people's details or DNA matches without consent.
• Hiding conflicts to make the story neat instead of writing down your reasoning.
• Letting an AI tool fill gaps. It can invent records and citations; open every source yourself.